Wangai & 2 others v Cabinet Secretary, Ministry of Information, Communications and The Digital Economy & 2 others; National Assembly (Interested Party); Ideate Tech Policy Africa Limited (itpa) (Amicus Curiae) (Petition E003 of 2026) [2026] KEHC 5690 (KLR) (30 April 2026) (Ruling)

KER188/2026 REPUBLIC OF KENYA IN THE HIGH REPUBLIC OF KENYA IN THE HIGH COURT OF KENYA AT KERUGOY A CONSTITUTIONAL & HUMAN RIGHTS DIVISION HCCHRPET E003 OF 2026 IN THE MATTER OF: THE ABSENCE OF A COMPREHENSIVE ARTIFICIAL INTELLIGENCE POLICY, LEGISLATION AND REGULATORY FRAMEWORK IN KENYA AND THE RESULTANT THREATS TO DIGITAL RIGHTS AND FUNDAMENTAL FREEDOMS AND IN THE MATTER OF: THE ENFORCEMENT OF THE CONSTITUTION OF KENYA, 2010 AND IN THE MATTER OF: ARTICLE 10 OF THE CONSTITUTION OF KENYA ON NATIONAL VALUES & PRINCIPLES OF GOVERNANCE AND IN THE MATTER OF: ARTICLE 232 OF THE CONSTITUTION OF KENYA ON THE PRINCIPLES OF PUBLIC SERVICE AND IN THE MATTER OF: CHAPTER 6 OF THE CONSTITUTION OF KENYA ON LEADERSHIP AND INTEGRITY AND IN THE MATTER OF: THE DATA PROTECTION ACT, 2019 AND IN THE MATTER OF: THE ACCESS TO INFORMATION ACT, 2016 AND IN THE MATTER OF: THE FAIR ADMINISTRATIVE ACTION ACT, 2015 AND IN THE MATTER OF: THE CONSUMER PROTECTION ACT AND IN THE MATTER OF: THE EMPLOYMENT ACT AND IN THE MATTER OF: THE KENYA INFORMATION AND COMMUNICATIONS ACT AND IN THE MATTER OF: THE DOCTRINE OF LEGITIMATE EXPECTATIONS BETWEEN JOHN WANGAI…………………............................................................................................. 1ST PETITIONER PETER AGORO…………….................................................................................................... 2ND PETITIONER 1 KER188/2026 ANTONY MANYARA…………………………………......................................……………3RD PETITIONER THE CABINET SECRETARY, MINISTRY OF INFORMATION, COMMUNICATIONS VERSUS AND THE DIGITAL ECONOMY...........................................................................................1ST RESPONDENT THE PRINCIPAL SECRETARY, STATE DEPARTMENT FOR ICT & DIGITAL ECONOMY……................................……2ND RESPONDENT THE HON. ATTORNEY GENERAL..................................................................................... 3RD RESPONDENT AND NATIONAL ASSEMBLY…………….…..........................................................................INTERESTED PARTY AND IDEATE TECH POLICY AFRICA LIMITED (ITPA) ……................................……………..AMICUS CURIAE Background RULING [1] The Petitioners moved the Court for various substantive reliefs in the Petition dated 4/2/2026 as follows: “F. PRAYERS SOUGHT The Petitioners hereby pray to the Honourable Court to give the following orders as follows. i. A DECLARATION be and is hereby issued that the failure and/or unreasonable delay by the 1st and 2nd Respondents to develop, enact and/or implement a comprehensive artificial intelligence policy, legislative framework and regulatory system violates and threatens Articles 1, 2, 3, 10, 19, 20, 21, 22, 27, 28, 31, 33, 34, 35, 38, 40(5), 41, 43, 46, 47, 48, 50, 53, 54, 55, 56, 57 and all other applicable provisions of the Constitution of Kenya 2010, and is therefore unconstitutional. ii. A DECLARATION that the failure and/or unreasonable delay by the 1st and 2nd Respondents to develop, enact and/or implement a comprehensive artificial intelligence policy, legislative framework and regulatory system constitutes a violation of their constitutional positive obligations to respect, protect and promote fundamental rights and freedoms. iii. A DECLARATION be and is hereby issued that the deployment of high-risk artificial intelligence systems by the 1st and 2nd Respondents without adequate legal framework violates and threatens the Constitution of Kenya 2010. 2 KER188/2026 iv. AN ORDER OF MANDAMUS be and is hereby issued compelling the 1st and 2nd Respondents, within ninety (90) days, to develop and publish a comprehensive draft National Artificial Intelligence Policy Framework. v. AN ORDER OF MANDAMUS be and is hereby issued compelling the 1st and 2nd Respondents, within ninety (90) days, to table before the Interested Party draft legislation establishing a comprehensive AI governance framework. vi. ANY SUCH OTHER OR FURTHER ORDERS as this Honourable Court may deem just and expedient in the circumstances to remedy the aforesaid violation. vii. COSTS of this petition to be granted to the Petitioners.” [2] The backbone primary facts relied on by the petitioners are set out in part B of the Petition as follows: “B. FACTS RELIED UPON 1. Kenya currently lacks a comprehensive, AI-specific policy, legislation and regulatory framework despite the rapid and extensive deployment of artificial intelligence systems across both public and private sectors in Kenya. 2. Artificial intelligence systems are being deployed and actively used in Kenya by government agencies, financial institutions, educational institutions, healthcare providers, employers, law enforcement agencies, digital platforms and other entities. 3. The deployment and use of the AI systems is occurring without adequate legal safeguards, mandatory transparency requirements, algorithmic accountability mechanisms, algorithmic impact assessments, human rights due diligence, meaningful human oversight, explainability obligations, or effective complaint and redress procedures. 4. The existing legal and regulatory framework in Kenya does not adequately address the specific risks, challenges and rights implications posed by artificial intelligence systems. 5. The Data Protection Act 2019, while establishing important data protection principles and rights, does not contain AI-specific provisions on algorithmic transparency, explainability of automated decisions, mandatory algorithmic impact assessments for high-risk AI systems, specific protections against algorithmic discrimination, or governance frameworks for AI training data and AI-generated outputs. 6. The absence of a comprehensive AI regulatory framework creates present and imminent threats to fundamental rights and freedoms guaranteed under the Constitution of Kenya 2010, including rights to privacy, dignity, equality and freedom from discrimination, freedom of expression and media freedom, access to information, political rights and free and fair elections, property and intellectual property rights, fair labour practices, socio-economic rights including health, education and social security, consumer rights, fair administrative action, and the rights of vulnerable and marginalised groups. 7. AI-powered surveillance systems are being deployed and/or are reasonably likely to be deployed in Kenya by both state and non-state actors without adequate legal frameworks governing their use. 8. Automated decision-making and AI-driven scoring systems are being used and/or are reasonably foreseeable in contexts including employment recruitment, workplace performance 3 KER188/2026 monitoring, credit scoring and loan approval, access to financial services and micro-finance, insurance underwriting and premium calculation, law enforcement risk assessment and predictive policing without transparency, explainability or effective appeal mechanisms. 9. These algorithmic systems create serious risks of discrimination particularly where AI systems are trained on biased historical data, rely on proxies for protected characteristics, or lack adequate testing and auditing for fairness and non-discrimination. 10. Algorithmic content curation, recommendation systems and automated content moderation employed by digital platforms and social media services are shaping political discourse, news consumption and access to information in Kenya, with potential to amplify misinformation and hate speech, suppress legitimate political expression and dissent, manipulate public opinion, and undermine media pluralism and freedom of expression. 11. Deepfake technology and synthetic media, including AI-generated video, audio and images that convincingly impersonate real persons, are being developed and disseminated globally and are already accessible to actors in Kenya, posing imminent risks of election disinformation and manipulation of electoral processes through fabricated videos or audio of political candidates, political warfare and erosion of public trust in authentic media, identity theft, impersonation and fraud targeting individuals, financial institutions and government agencies, gender-based violence including nonconsensual intimate images and sexualised deepfakes and reputational harm and violations of the right to dignity. 12. Kenya is approaching a general election cycle and there exist serious and imminent risks that AI tools could be used to interfere with free and fair elections, manipulate voters, suppress turnout, undermine confidence in electoral institutions, and violate political rights. 13. Rapid AI-driven automation and job displacement is occurring across various sectors in Kenya without adequate legal frameworks for worker protection thereby threatening the right to fair labour practices and socio-economic rights. 14. AI-powered algorithmic management systems are being deployed in workplaces without transparency, public participation, due process safeguards or protection from unfair or arbitrary algorithmic decisions, undermining fair labour practices and worker dignity. 15. Children in Kenya are exposed to AI systems through educational technology platforms, social media and digital services, online gaming and content recommendation algorithms without adequate protections for informed parental consent, age-appropriate design and content, protection from manipulative and addictive features, algorithmic profiling, exposure to harmful and inappropriate content thereby violating children's rights. 16. The use of AI systems in educational institutions is occurring without adequate regulatory framework ensuring fairness, transparency, human oversight, appeal mechanisms, or protection against algorithmic bias and discrimination. 17. Academic integrity in Kenya's education system is threatened by widespread availability of generative AI tools capable of producing essays, assignments and research papers, creating risks of undisclosed AI-generated academic work, plagiarism, and erosion of 4 KER188/2026 learning outcomes, all in the absence of clear national policies, institutional guidelines, or appropriate regulatory responses. 18. AI systems are being deployed in government decision-making and public administration without adequate safeguards, transparency and explainability of algorithmic decisions, effective appeal and review mechanisms, protection from algorithmic errors and bias and fair administrative action. 19. Persons with disabilities, older persons, youth, minorities, marginalised communities and other vulnerable groups face heightened risks from AI systems contrary to the Constitution which require special measures and consideration for vulnerable and marginalised groups. 20. Unregulated AI-enabled consumer products and services are being marketed and sold in Kenya without clear legal frameworks for product safety and quality standards, liability for AI-related harms and malfunctions, transparency about AI capabilities and limitations, disclosure of AI use, consumer protection from deceptive AI practices and effective redress mechanisms, thereby violating consumer rights. 21. Kenyan creators and artists are having their creative works copied and used without authorisation to train commercial AI systems, and subsequently having AI-generated outputs directly compete with their original works in the market, all without consent, compensation or legal protections, thereby violating intellectual property rights. 22. There exist no clear legal frameworks in Kenya governing liability for harms caused by AI systems, requirements for algorithmic transparency and explainability, mandatory impact assessments for high-risk AI applications, monitoring and auditing of deployed AI systems, incident reporting and disclosure obligations, accessible complaint and redress mechanisms for persons harmed by AI systems or enforcement mechanisms and sanctions for AI-related violations. 23. The Respondents have failed to establish regulatory institutions with adequate expertise, capacity, independence and authority to oversee AI systems, conduct investigations, enforce standards, mediate disputes, or provide guidance on AI governance, leaving persons affected by AI systems without effective access to justice or remedies. 24. This Honourable Court in the Huduma Namba case emphasised that intrusive digital identification systems must be preceded by adequate legislation and regulatory frameworks to safeguard privacy, prevent data breaches, address risks of misuse, and ensure non- discrimination. 26. International frameworks and best practices emphasise the importance of human rights and dignity-centred AI governance, transparency and explainability, algorithmic fairness and non-discrimination, accountability and oversight, safety and security, privacy and data protection, multi-stakeholder governance and public participation. 27. The rapid pace of AI development and deployment, combined with the absence of adequate regulation, creates an urgent and growing gap between technological capabilities and legal safeguards, exposing Kenyans to increasing risks of rights violations that may become irreparable if not addressed through timely regulatory intervention. 5 KER188/2026 28. The Respondents have failed and/or unreasonably delayed in developing, publishing and implementing a comprehensive AI policy, legislative framework and regulatory system despite the growing deployment of AI systems in Kenya, the clear and present threats to constitutional rights, their constitutional obligations to respect, protect, promote and fulfill rights, their obligations under the national values and principles of governance including transparency, accountability and good governance; and international and regional calls for AI governance frameworks.” [3] The alleged adverse effect of the default and the result on various aspects of life are characterised as violations of the constitutional rights and freedoms as set out in part C of the Petition as follows: “C. NATURE OF INJURY Violations and threats to the right to privacy, as provided for under Article 31 of the Constitution through unregulated AI-enabled surveillance, facial recognition, biometric processing, invasive collection and processing of personal data, profiling, tracking and monitoring of individuals and groups without adequate legal basis, transparency, safeguards, oversight or redress. Violations and threats to the right to equality and freedom from discrimination, as stipulated under Article 27 of the Constitution through algorithmic discrimination, bias and exclusion in employment, hiring, credit-scoring, access to financial services, insurance, education, healthcare, law enforcement, access to government services and social protection, and other critical domains affecting life chances and fundamental rights. Violations and threats to inherent dignity and the right to have dignity respected and protected, pursuant to Article 28 of the Constitution through deepfakes, sexualised synthetic content and non-consensual intimate images. Violations and threats to freedom of expression as provided for under Article 33 and media freedom as stipulated under Article 34 of the Constitution through opaque, arbitrary and sometimes discriminatory algorithmic content moderation, censorship and suppression of speech. Violations and threats to the right to access information pursuant to Article 35 of the Constitution through lack of transparency regarding AI use in government and private sector, algorithmic opacity preventing meaningful understanding of how decisions are made, denial of explainability and accountability, and information asymmetries between those who deploy AI and those affected by it. Violations and threats to political rights, as provided for under Article 38 of the Constitution including rights to free, fair and regular elections, voter registration, and political participation through AI-driven electoral disinformation and deepfakes, manipulation of political discourse and public opinion, algorithmic micro-targeting and manipulation of voters, and interference with the integrity of democratic processes and institutions. Violations and threats to the right to property including intellectual property rights, as stipulated under Article 40 of the Constitution through unauthorised use, copying 6 KER188/2026 and exploitation of creative works for AI training and generation of competing outputs without consent, attribution or fair compensation to creators. Violations and threats to labour rights, as provided for under Article 41 of the Constitution including rights to fair labour practices and fair remuneration through AIdriven job displacement without adequate protections, workplace surveillance without safeguards for dignity and fairness, and concentration of economic gains from AI among technology companies and capital owners rather than workers and society. Violations and threats to socio-economic rights, as provided for under Article 43 of the Constitution including rights to health, education, social security, food and clean water through deployment of AI systems in these domains without adequate transparency, safeguards, accountability, human oversight or mechanisms to prevent and remedy discriminatory exclusion from essential services. Violations and threats to consumer rights, as provided for under Article 46 of the Constitution through deployment of AI-enabled products and services without adequate safety standards, quality assurance, transparency regarding AI use, effective remedies when AI systems cause consumer harm, or clear liability frameworks. Violations and threats to the right to fair administrative action, as provided for under Article 47 of the Constitution through automated or algorithmically-influenced government decisions without adequate transparency, explainability, lawful procedure, human oversight, reasons for decisions, or effective rights to appeal and review. Violations and threats to the right to access justice, as provided for under Article 48 and the right to fair hearing as enshrined under Article 50 of the Constitution through lack of effective remedies, accessible complaint mechanisms, clear liability frameworks and judicial capacity to address AI-related harms. Violations and threats to the rights of children, as provided for under Article 53 of the Constitution through exposure to harmful AI-generated content, algorithmic manipulation, profiling and targeting of children, automated decisions affecting educational opportunities, and absence of child-specific AI safeguards and ageappropriate digital rights education. Violations and threats to the rights of persons with disabilities, as provided for under Article 54 of the Constitution through algorithmic discrimination based on disability, lack of accessibility in AI systems and exclusion from AI governance processes. Violations and threats to the rights of youth, as provided for under Article 55 of the Constitution through AI-driven job displacement disproportionately affecting youth employment prospects, algorithmic systems affecting access to education and economic opportunities, and exclusion from meaningful participation in AI policy affecting their futures. Violations and threats to the rights of minorities and 7 KER188/2026 marginalised groups, as provided for under Article 56 of the Constitution through algorithmic profiling and discrimination based on ethnicity, religion, language or socio-economic status and lack of participation in AI governance. Violations and threats to the rights of older persons, as provided for under Article 57 of the Constitution through age-based algorithmic discrimination and underrepresentation in AI governance. Undermining of the rule of law, democratic governance, accountability, transparency, integrity and other foundational values and principles enshrined in Articles 10 and 73 of the Constitution through deployment of opaque, unaccountable and inadequately governed AI systems. Erosion of public trust and confidence in government institutions, democratic processes, electoral integrity, authentic information, digital evidence, and the capacity of law and the Constitution to protect rights in the digital age. Denial of effective enforcement of the Constitution, access to justice and remedies for persons harmed by AI systems due to absence of clear frameworks for accountability, liability, standing, causes of action and appropriate relief in AI contexts.” [4] Consequently, specific substantive reliefs are sought as follows: “F. PRAYERS SOUGHT The Petitioners hereby pray to the Honourable Court to give the following orders as follows. i. A DECLARATION be and is hereby issued that the failure and/or unreasonable delay by the 1st and 2nd Respondents to develop, enact and/or implement a comprehensive artificial intelligence policy, legislative framework and regulatory system violates and threatens Articles 1, 2, 3, 10, 19, 20, 21, 22, 27, 28, 31, 33, 34, 35, 38, 40(5), 41, 43, 46, 47, 48, 50, 53, 54, 55, 56, 57 and all other applicable provisions of the Constitution of Kenya 2010, and is therefore unconstitutional. ii. A DECLARATION that the failure and/or unreasonable delay by the 1st and 2nd Respondents to develop, enact and/or implement a comprehensive artificial intelligence policy, legislative framework and regulatory system constitutes a violation of their constitutional positive obligations to respect, protect and promote fundamental rights and freedoms. iii. A DECLARATION be and is hereby issued that the deployment of high-risk artificial intelligence systems by the 1st and 2nd Respondents without adequate legal framework violates and threatens the Constitution of Kenya 2010. iv. AN ORDER OF MANDAMUS be and is hereby issued compelling the 1st and 2nd Respondents, within ninety (90) days, to develop and publish a comprehensive draft National Artificial Intelligence Policy Framework. 8 KER188/2026 v. AN ORDER OF MANDAMUS be and is hereby issued compelling the 1st and 2nd Respondents, within ninety (90) days, to table before the Interested Party draft legislation establishing a comprehensive AI governance framework. vi. ANY SUCH OTHER OR FURTHER ORDERS as this Honourable Court may deem just and expedient in the circumstances to remedy the aforesaid violation. vii. COSTS of this petition to be granted to the Petitioners.” Application for conservatory orders [5] The Petitioners have sought conservatory orders pending the hearing and determination of the Petition by a Notice of Motion dated 4/2/2026 as follows: “ORDERS: - i. THAT this Application be certified as urgent and service thereof be dispensed with in the first instant. ii. THAT pending the hearing and determination of this Application inter parties, this Honourable Court be pleased to issue a Conservatory Order restraining the 1st and 2nd Respondents, whether by themselves, their officers, servants, agents, or any person acting on their behalf, from deploying, authorizing and/or operationalizing the use of high-risk artificial intelligence systems. iii. THAT this Honourable Court be pleased to order that this Application and the Petition be heard and determined on a priority basis given the urgency, public importance and constitutional significance of the matters raised. iv. THAT for reasons to be recorded service of this application be dispensed with and the present application be heard ex parte in the first instance in respect of prayer 2 hereof. v. THAT this Honourable Court be at liberty to issue any further orders in the interests of justice. vi. THAT the Honourable Court be pleased to give directions relating to the hearing and determination of this Petition. vii. THAT the costs of this Application be provided for.” [6] The grounds of the application were set out in the application as follows: “G rounds that: a) This Application is urgent and ought to be heard on priority basis. b) Kenya is experiencing rapid and widespread deployment of AI systems without adequate legal and regulatory frameworks, creating imminent threats to fundamental rights and freedoms. c) AI systems are being deployed without adequate legal basis, transparency, human oversight or effective safeguards, in violation of constitutional rights. d) Kenyans are experiencing and imminently threatened with violations of their rights to privacy, equality, non-discrimination, dignity, fair administrative action, freedom of 9 KER188/2026 expression, political participation, labour rights, consumer protection and other constitutional guarantees due to unregulated AI deployment. e) The 1st and 2nd Respondents have failed to fulfil their positive constitutional obligations as enshrined under Article 21 of the Constitution to respect, protect, promote and fulfil fundamental rights by establishing adequate legal and institutional frameworks before allowing deployment of technologies that pose systemic risks to such rights. f) This Honourable Court held in Nubian Rights Forum & 2 Others v Attorney General & 6 Others that robust legal and regulatory safeguards must precede deployment of intrusive digital systems. g) International and regional human rights bodies have emphasised the urgent need for comprehensive AI governance frameworks compatible with human rights obligations. h) Kenya is approaching general elections in 2027 and the lack of AI governance creates grave and imminent risks of electoral manipulation through deepfakes, disinformation, algorithmic interference with political discourse and other threats to free and fair elections. i) Automated and AI-assisted government decision-making is being implemented without adequate human oversight, transparency, explainability, fairness safeguards or effective review and appeal mechanisms. j) Children, persons with disabilities, older persons, marginalised communities and other vulnerable groups face heightened risks from AI systems without adequate protections tailored to their specific vulnerabilities. k) Consumers face risks from unregulated AI-powered products and services where there is no clear liability framework, safety standards, quality assurance, transparency requirements or effective redress mechanisms, undermining consumer rights. l) The academic integrity and education system are threatened by undisclosed AI generated work, biased AI assessment systems, plagiarism and copyright violations, in the absence of a national framework, thereby undermining the integrity of education and the right to education. m) Kenyan creators, artists, writers, musicians and other rights holders are having their creative works and intellectual property used for AI training and AI generated outputs without consent, fair compensation or adequate safeguards, in violation of their intellectual property rights. n) Algorithmic profiling, automated decision-making and AI systems are being used in employment, hiring, credit-scoring, micro-finance, insurance, access to public services and law enforcement without transparency, explainability, fairness safeguards or effective appeal mechanisms, thereby creating risks of discrimination, exclusion and denial of rights. o) Rapid AI deployment is causing and/or will cause job displacement and labour market disruption without adequate worker protections, social safety nets, retraining or 10 KER188/2026 upskilling programmes thereby undermining the right to fair labour practices and threatening socio-economic rights. p) Persons harmed by AI-mediated decisions, algorithmic discrimination, biased automated systems or other AI-related violations face significant barriers to access to justice due to legislative and institutional gaps, lack of transparency in AI systems, absence of explainability requirements, inadequate liability frameworks and lack of specialised expertise in the judiciary and regulatory bodies. q) Last year, the 1st and 2nd Respondents launched the National AI Strategy 2025 – 2030 which outlines policy goals for AI adoption, ethics, and infrastructure. r) Currently, no dedicated AI legislation exists, as existing laws provide only partial coverage and fall short of addressing the unique and systemic risks posed by AI systems. s) The Computer Misuse and Cybercrimes (Amendment) Act, 2025 which addresses cyber-enabled crimes and which could apply to some AI misuse like AI-generated malicious content or deepfake fraud, was partially suspended and therefore not fully operational. t) The Petitioners satisfies the twin requirements for grant of Conservatory Orders, the arguability of the Petition and the prejudice unless the Conservatory orders are issued, per Centre for Rights Education and Awareness (CREAW) & 7 Others –v- Attorney General [2011] eKLR, that; “… a party seeking a conservatory order only requires to demonstrate that he has a prima facie case with a likelihood of success and that unless the court grants the conservatory order there is real danger that he will suffer prejudice as a result of the violation or threatened violation of the Constitution.” (Emphasis added). u) The applicable principles for the grant of conservatory orders were also detailed by the Hon Onguto J. in the precedent set by Board of Management of Uhuru Secondary School v. City County Director of Education & 2 Others [2015] eKLR which inter alia include; an arguable prima facie case with a likelihood of success and that in the absence of the conservatory orders the petitioner(s) is likely to suffer prejudice, if an interim conservatory order is not granted, the petition or its substratum will be rendered nugatory, public interest and relevant material facts. v) It is no doubt in the interests of the fair administration of justice, legitimate expectation and the rule of law that the application filed herewith is admitted urgently for the urgent reliefs thereof to prevent the otherwise devastating violation of fundamental rights of Kenyan citizens guaranteed by the Constitution of Kenya 2010.” [7] The supporting affidavit of the 2nd Petitioner sworn on the 4/2/2026 was in the same text as the grounds of the application above, as follows: “SUPPORTING AFFIDAVIT 11 KER188/2026 I, PETER AGORO, a patriotic Kenya citizen and of Post Office Box Number 17586 - 00100, NAIROBI, do hereby make oath and state as follows THAT: 1. I am the Petitioner, conversant with the facts in issue and competent to swear this affidavit in this public interest litigation. 2. Kenya is experiencing rapid and widespread deployment of AI systems without adequate legal and regulatory frameworks, creating imminent threats to fundamental rights and freedoms. 3. AI systems are being deployed without adequate legal basis, transparency, human oversight or effective safeguards, in violation of constitutional rights. 4. Kenyans are experiencing and imminently threatened with violations of their rights to privacy, equality, non-discrimination, dignity, fair administrative action, freedom of expression, political participation, labour rights, consumer protection and other constitutional guarantees due to unregulated AI deployment. 5. The 1st and 2nd Respondents have failed to fulfil their positive constitutional obligations as enshrined under Article 21 of the Constitution to respect, protect, promote and fulfil fundamental rights by establishing adequate legal and institutional frameworks before allowing deployment of technologies that pose systemic risks to such rights. 6. This Honourable Court held in Nubian Rights Forum & 2 Others v Attorney General & 6 Others that robust legal and regulatory safeguards must precede deployment of intrusive digital systems. 7. International and regional human rights bodies have emphasised the urgent need for comprehensive AI governance frameworks compatible with human rights obligations. 8. Kenya is approaching general elections in 2027 and the lack of AI governance creates grave and imminent risks of electoral manipulation through deepfakes, disinformation, algorithmic interference with political discourse and other threats to free and fair elections. 9. Automated and AI-assisted government decision-making is being implemented without adequate human oversight, transparency, explainability, fairness safeguards or effective review and appeal mechanisms. 10. Children, persons with disabilities, older persons, marginalised communities and other vulnerable groups face heightened risks from AI systems without adequate protections tailored to their specific vulnerabilities. 11. Consumers face risks from unregulated AI-powered products and services where there is no clear liability framework, safety standards, quality assurance, transparency requirements or effective redress mechanisms, undermining consumer rights. 12. The academic integrity and education system are threatened by undisclosed AI generated work, biased AI assessment systems, plagiarism and copyright violations, 12 KER188/2026 in the absence of a national framework, thereby undermining the integrity of education and the right to education. 13. Kenyan creators, artists, writers, musicians and other rights holders are having their creative works and intellectual property used for AI training and AI generated outputs without consent, fair compensation or adequate safeguards, in violation of their intellectual property rights. 14. Algorithmic profiling, automated decision-making and AI systems are being used in employment, hiring, credit-scoring, micro-finance, insurance, access to public services and law enforcement without transparency, explainability, fairness safeguards or effective appeal mechanisms, thereby creating risks of discrimination, exclusion and denial of rights. 15. Rapid AI deployment is causing and/or will cause job displacement and labour market disruption without adequate worker protections, social safety nets, retraining or upskilling programmes thereby undermining the right to fair labour practices and threatening socio-economic rights. 16. Persons harmed by AI-mediated decisions, algorithmic discrimination, biased automated systems or other AI-related violations face significant barriers to access to justice due to legislative and institutional gaps, lack of transparency in AI systems, absence of explainability requirements, inadequate liability frameworks and lack of specialised expertise in the judiciary and regulatory bodies. 17. Last year, the 1st and 2nd Respondents launched the National AI Strategy 2025 – 2030 which outlines policy goals for AI adoption, ethics, and infrastructure. 18. Currently, no dedicated AI legislation exists, as existing laws provide only partial coverage and fall short of addressing the unique and systemic risks posed by AI systems. 19. The Computer Misuse and Cybercrimes (Amendment) Act, 2025 which addresses cyber-enabled crimes and which could apply to some AI misuse like AI-generated malicious content or deepfake fraud, was partially suspended and therefore not fully operational. 20. The present Application and Petition filed herewith are accordingly extremely urgent, as they seek to defend the Constitution, and public interest. 21. It is therefore critical that the Honourable Court intervenes as prayed for in the present Application and the Petition filed herewith, to forestall the otherwise devastating consequences that the 1st and 2nd Respondents actions will occasion to public interest, fundamental rights and the Rule of Law. 22. I make this Affidavit in support of both the Application and Petition filed herewith. 23. The matters stated hereinabove are true and are within my own knowledge save as to matter deposed to on information and belief, the whereof and grounds whereupon are disclosed.” 13 [8] The Respondents have filed Grounds of Opposition dated 1/4/2026 to the application dated 4/2/2026 as follows: KER188/2026 “1st. 2nd and 3rd RESPONDENTS' GROUNDS OF OPPOSITION 1. THAT the applicants have failed to establish a prima facie case with a likelihood of success to warrant grant of conservatory orders as illustrated in Mrao Ltd vs. First American Bank of Kenya & 2 others [2003] KLR 125, where a prima facie case was described as: "In civil case, a prima facie case is a case in which on the material presented to the Court a tribunal properly directing itself will conclude that there exists a right which has apparently been infringed by the opposite party as to call for an explanation or rebuttal from the latter”. 2. THAT the applicant has yet to demonstrate any real or imminent prejudice to warrant grant of conservatory orders. 3. THAT the granting the orders sought would undermine innovation, public service delivery and national development contrary to public interest. 4. THAT the application is premised on speculative and generalized allegations unsupported by evidence. 5. THAT the orders sought unjustifiably interfere with the constitutional and statutory mandate of the respondents in policy formulation and implementation. Intervention of the Court would undermine the principle of separation of powers as policy formulation falls within the mandate of the executive. 6. THAT the orders sought are vague and incapable of practical enforcement, particularly the blanket restraint on high risk artificial intelligence systems. 7. THAT the applicant have failed to demonstrate a complete absence of legal or regulatory mechanisms governing the issues raised in their application. 8. THAT the application is misconceived, mischievous and gross abuse of the Court process. 9. THAT the application is devoid of merit and ought to be dismissed with costs. J.M. KIONGO, PRINCIPAL STATE COUNSEL FOR: HON. ATTORNEY GENERAL.” [9] The other parties did not file Grounds of Opposition/Replying Affidavits on the application. Submissions [10] The parties then filed submissions on the application and ruling was reserved. Counsel for the Interested Party/National Assembly and Counsel for the Amicus expressed intention not to make any submissions at the interlocutory stage of the Notice of Motion for conservatory orders and to participate only at the full hearing of the Petition. [11] The Petitioner who indicated that he was not acting on behalf of the co-petitioners submitted by Written Submissions dated 17th March 2026 on the dangers and need to positively respond to the AI threat by halting its deployment ad use, principally, as follows: 14 KER188/2026 “4. SUBSTANTIVE ARGUMENTS 4.1 The Petitioners Have Established a Prima Facie Case with a Likelihood of Success 4.1.1 Overview of the Prima Facie Case The Petitioners have established a strong prima facie case that the absence of a comprehensive AI policy, legislation and regulatory framework, coupled with ongoing unregulated deployment of high-risk AI systems, constitutes a violation of multiple fundamental rights and freedoms guaranteed under the Constitution of Kenya 2010. The prima facie case rests on three interrelated pillars: 1. First, artificial intelligence systems are being deployed across Kenya in contexts affecting fundamental rights including government services, financial services, employment, education, law enforcement, healthcare, and digital platforms without adequate legal safeguards, transparency requirements, accountability mechanisms, or effective remedies; 2. Second, the existing legal framework in Kenya including the Data Protection Act 2019, the Access to Information Act 2016, the Fair Administrative Action Act 2015, the Consumer Protection Act 2012, and the partially suspended Computer Misuse and Cybercrimes (Amendment) Act 2025 does not contain AI-specific provisions adequate to address the unique risks posed by AI systems; 3. Third, the 1st and 2nd Respondents have failed to fulfil their positive constitutional obligations under Articles 20 and 21 to respect, protect, promote and fulfil fundamental rights by establishing adequate legal and institutional frameworks before allowing deployment of technologies that pose systemic risks to such rights. 4.1.2 Unregulated High-Risk AI Deployment Violates Fundamental Rights Violations of the Right to Privacy (Article 31 of the Constitution of Kenya 2010) AI- powered surveillance systems, including facial recognition, biometric data processing, location tracking and profiling technologies, are being deployed by both state and non-state actors without clear legal authority, adequate data protection safeguards, human rights impact assessments or mechanisms for oversight and redress. These AI systems are being deployed without adequate safeguards, oversight, accountability mechanisms, transparency requirements, impact assessments or effective remedies for persons adversely affected. The right to privacy is not merely a right to be left alone, but encompasses informational self- determination, freedom from invasive profiling and surveillance, and protection Page 9 against the use of personal data in ways that may harm individuals or create chilling effects on the exercise of other rights. AI-enabled surveillance and profiling pose unprecedented threats to privacy because they enable mass, indiscriminate, continuous and automated monitoring and analysis of individuals' movements, behaviours, associations and communications. 15 KER188/2026 Violations of the Right to Equality and Non-Discrimination (Article 27 of the Constitution of Kenya 2010) Algorithmic profiling, automated decision-making and AI systems are being used in employment, hiring, credit-scoring, micro-finance, insurance, access to public services, law enforcement, criminal justice and social protection without transparency, explainability, fairness safeguards or effective appeal mechanisms. Algorithmic discrimination occurs when AI systems trained on biased historical data, or relying on proxies for protected characteristics such as race, gender, ethnicity, or socioeconomic status, systematically disadvantage certain groups. Unlike human discrimination, algorithmic discrimination operates at scale, is often invisible to those affected, and is difficult to challenge due to lack of transparency and explainability in AI systems. Threats to Human Dignity (Article 28 of the Constitution of Kenya 2010) Deepfakes, synthetic media and generative AI technologies pose imminent and serious risks of election disinformation, identity theft, impersonation, fraud, gender-based abuse through sexualized deepfakes, reputational harm and erosion of public trust. The absence of any legal framework to prevent, detect, attribute or remedy such harms leaves individuals vulnerable to profound violations of dignity through non- consensual intimate images, impersonation, and manipulation of their identity and likeness. Threats to Freedom of Expression and Media Freedom (Articles 33 and 34 of the Constitution of Kenya 2010) AI-driven content curation, recommender systems and algorithmic content moderation on digital platforms are shaping public discourse, political expression and access to information in opaque ways that may distort democratic debate, amplify hate speech and misinformation, censor legitimate expression and undermine freedom of expression and media freedom. Unlike traditional editorial gatekeeping, algorithmic content curation operates through opaque automated systems that can amplify or suppress content based on engagement metrics rather than editorial judgment, creating echo chambers, filter bubbles and distorted information environments that undermine informed public discourse. Threats to Political Rights and Electoral Integrity (Article 38 of the Constitution of Kenya 2010) The forthcoming general elections scheduled for August 2027 face heightened risks from AI-driven disinformation, deepfakes, synthetic media, algorithmic manipulation of political discourse and electoral interference, which threaten the integrity of democratic processes and the constitutional right to free and fair elections. The 2027 elections are approaching and Kenya has no legal framework to prevent or remedy AI-enabled electoral manipulation. This is not a hypothetical threat as deepfakes technology and generative AI have already been used to manipulate elections in other jurisdictions, and such technologies are readily accessible in Kenya. 16 KER188/2026 Violations of Labour Rights (Article 41 of the Constitution of Kenya 2010) Rapid AI deployment is causing and will cause job displacement, deskilling and labour market disruption without adequate worker protections, social safety nets, social dialogue, retraining or upskilling programmes, or just transition frameworks, thereby undermining the right to fair labour practices and threatening socio-economic rights. Violations of Fair Administrative Action (Article 47 of the Constitution of Kenya 2010) When government decisions that affect individuals’, rights are made or substantially influenced by AI systems, the right to fair administrative action requires that individuals be able to understand how decisions were made, challenge erroneous or biased decisions, and obtain meaningful review. None of these safeguards currently exist in Kenya's AI deployment. 4.1.3 Existing Legal Framework is Inadequate The Petitioners submit that the existing legal framework in Kenya does not contain AI specific provisions adequate to address the unique risks posed by AI systems: The Data Protection Act 2019 The Act's provisions on automated decision-making are minimal and do not address the specific challenges posed by modern AI systems including machine learning, neural networks, generative AI and other technologies that operate in ways fundamentally different from traditional automated processing. The Computer Misuse and Cybercrimes (Amendment) Act 2025 This Act addresses cyber-enabled crimes and which could apply to some AI misuse like AI-generated malicious content or deepfakes fraud, was partially suspended and therefore not fully operational. Even if fully operational, this Act addresses criminal misuse of AI but does not establish a comprehensive governance framework for AI deployment, oversight and accountability. The National AI Strategy 2025-2030 Last year, the 1st and 2nd Respondents launched the National AI Strategy 2025 – 2030 which outlines policy goals for AI adoption, ethics, and infrastructure. However, a non- binding strategy document is fundamentally insufficient to protect constitutional rights. It does not: (p) Create legally enforceable rights or obligations; (q) Establish regulatory institutions with enforcement powers; (r) Provide mechanisms for oversight, investigation or sanction; (s) Create rights of action or remedies for persons harmed by AI systems; (t) Mandate transparency, impact assessments, or other concrete safeguards. A strategy document, no matter how well-intentioned, cannot substitute for legislation that creates enforceable legal obligations, establishes regulatory institutions, provides for oversight and accountability, and creates effective remedies. 4.1.4 The Nubian Rights Forum Principle Applies with Greater Force to AI In Nubian Rights Forum & 2 Others v Attorney General & 6 Others [2020] eKLR, this Honourable Court held that the Government's plan to roll out the Huduma Namba digital identification system without adequate data protection legislation and regulatory framework violated constitutional rights. The Court emphasized that comprehensive legal and regulatory safeguards must be in 17 KER188/2026 place before deployment of intrusive digital systems. AI systems are not merely data processing tools but they are technologies that can make decisions, generate content, automate judgments, profile individuals, predict behaviour and shape information environments in ways that were impossible with earlier technologies. The risks they pose are qualitatively different and quantitatively greater than traditional digital systems. If a comprehensive legal framework was required before deploying a digital identification system, a fortiori such a framework is required before allowing widespread Page 13 deployment of AI systems that pose even greater and more diffuse risks to fundamental rights and democratic processes. 4.1.5 Respondents Have Failed Their Positive Constitutional Obligations Article 21 of the Constitution establishes that it is a fundamental duty of the State and every State organ to observe, respect, protect, promote and fulfil the rights and fundamental freedoms in the Bill of Rights. The obligation to protect rights requires the State to take measures to prevent violations by third parties, not merely to refrain from violations itself. The obligation to fulfil rights requires the State to take positive legislative, administrative and other measures to enable rights to be realized. By allowing unregulated deployment of high-risk AI systems without first establishing adequate legal safeguards, regulatory oversight and effective remedies, the Respondents have failed to protect Kenyan citizens from rights violations by AI deployers and have failed to fulfil their obligation to create the legislative and institutional framework necessary for rights to be realized in the AI era. Conclusion on Prima Facie Case: The Petitioners have established a strong prima facie case that unregulated high-risk AI deployment violates multiple fundamental rights, that existing law is inadequate, and that the Respondents have breached their positive constitutional obligations. This case is highly arguable and has a strong likelihood of success on the merits. 4.2 The Petitioners Will Suffer Irreparable Injury Without Conservatory Orders The second requirement for grant of conservatory orders is that the petitioners will suffer irreparable injury or prejudice if the orders are not granted. The Petitioners submit that the threatened and ongoing violations of constitutional rights caused by unregulated high-risk AI deployment constitute irreparable injury of the highest order. Page 14 4.2.1 Rights Violations Cannot Be Adequately Remedied by Damages Violations of fundamental rights such as privacy, dignity, equality, fair administrative action and political participation cannot be adequately compensated through monetary damages. Once AI systems have been used to profile individuals, make discriminatory decisions, manipulate political discourse, or deploy surveillance, the harm cannot be undone. Privacy lost cannot be restored. Dignity violated cannot be fully repaired. Elections manipulated cannot be re-run. 4.2.2 AI Systems Cannot Be Undeployed Once high-risk AI systems are deployed, their effects are difficult or impossible to reverse: (u) Biometric data collected by facial recognition or surveillance systems 18 KER188/2026 cannot be 'uncollected'; (v) Profiles created by algorithmic systems become embedded in decision-making and cannot be fully erased; (w) Discriminatory AI decisions that denied employment, credit, or services have lasting consequences on individuals' life prospects; (x) Deepfakes and synthetic media, once circulated, cannot be fully removed from the internet; (y) AI-manipulated electoral processes cannot be reversed—if the August 2027 elections are corrupted by AI-driven disinformation, the damage to democratic legitimacy will be irreparable. 4.2.3 Imminent Threat to 2027 Elections The threat to the integrity of the August 2027 elections is both imminent and irreparable. Kenya is approximately 18 months from a general election. AI tools capable of generating convincing deepfakes of political candidates, creating synthetic disinformation at scale, and manipulating online political discourse are readily available. Without a legal framework to prevent, detect and remedy AI-driven electoral manipulation, the constitutional right to free and fair elections under Article 38 faces unprecedented threat. Once an election has been corrupted by AI- enabled manipulation, the damage cannot be undone. The legitimacy of the elected government and public confidence in democratic institutions will be irreparably harmed. This is not speculation—AI-enabled electoral manipulation has already occurred in other jurisdictions. The United States, European Union, India, and multiple other democracies have experienced AI-generated deepfakes of political candidates and AI-driven disinformation campaigns. Kenya cannot afford to wait until after such manipulation has occurred to establish a regulatory framework. 4.2.4 Systemic Harm to Vulnerable Groups As detailed in the Petition and Certificate of Urgency, children, persons with disabilities, older persons, marginalised communities and other vulnerable groups face heightened risks from unregulated AI systems. These groups often lack the resources, technical literacy or access to justice necessary to protect themselves from AI-related harms. Once these groups are harmed by discriminatory algorithms, invasive profiling or manipulative AI systems, the damage to their rights and wellbeing cannot be fully remedied. 4.2.5 Erosion of Public Trust and Rule of Law The deployment of opaque, unaccountable AI systems by government and powerful private entities, in the absence of legal safeguards and oversight, erodes public trust in institutions, undermines the rule of law, and creates a sense that technology is beyond Page 16 democratic control. This erosion of trust and confidence in institutions is itself an irreparable harm that cannot be measured in monetary terms or remedied through compensation. Conclusion on Irreparable Harm: The Petitioners and the Kenyan public will suffer irreparable harm if high-risk AI systems continue to be deployed without adequate legal safeguards. Violations of fundamental rights, manipulation of democratic processes, and erosion of public 19 KER188/2026 trust cannot be adequately remedied through damages after the fact. Conservatory orders are necessary to prevent irreparable injury. 4.3 The Substratum of the Petition Will Be Rendered Nugatory Without Orders 4.3.1 The third consideration is whether the petition will be rendered nugatory without conservatory orders. The Petitioners submit that if high-risk AI systems continue to be deployed in an unregulated manner while this case proceeds, the very purpose of the litigation which is to secure constitutional rights before irreversible harm occurs, will be defeated. 4.3.2 The substratum of this Petition is the principle, established in Nubian Rights Forum / Huduma Number case, that comprehensive legal and regulatory frameworks must be in place before deployment of intrusive systems affecting fundamental rights. The Petition seeks declarations that the Respondents' failure to establish such frameworks before allowing AI deployment violates the Constitution, and seeks orders compelling the Respondents to develop and enact such frameworks. 4.3.3 If high-risk AI systems are allowed to proliferate during the pendency of this case, several consequences follow: First, the principle that safeguards must precede deployment becomes meaningless, if AI can be deployed freely while litigation challenging unregulated deployment proceeds, the litigation cannot secure the preventive protection that the Constitution requires; Second, AI systems that are deployed become embedded in institutional practices, create path dependencies, generate vested interests, and become politically and practically difficult to regulate or remove and thus any victory on the merits would be largely symbolic; Third, the constitutional rights violations that the Petition seeks to prevent will have already occurred, surveillance will have been conducted, discriminatory decisions will have been made, democratic processes will have been manipulated; Fourth, and most critically, if the August 2027 elections occur while AI remains unregulated, the electoral integrity concerns that are central to this Petition will be moot as the election will have passed, and any finding that unregulated AI posed threats to electoral integrity will provide no remedy. 4.3.4 The principle established in Board of Management of Uhuru Secondary School v County Director of Education & 2 Others [2015] eKLR is that if interim orders are not granted, the petition or its substratum will be rendered nugatory. This test is clearly satisfied here. Without conservatory orders, the preventive purpose of the Petition will be defeated, and the Petition will become an exercise in documenting violations that have already occurred rather than preventing them. Conclusion on Nugatory Petition: Without conservatory orders, the substratum of the Petition which is securing constitutional rights before irreversible harm occurs, will be defeated. The Petition will be rendered nugatory and an academic exercise if AI continues to be deployed without safeguards while this case proceeds. 4.4 Balance of Convenience and Public Interest Favour Grant of Orders The fourth consideration is 20 KER188/2026 the balance of convenience and public interest. The Petitioners submit that both overwhelmingly favour the grant of conservatory orders. 4.4.1 Public Interest in Constitutional Rights and Democratic Integrity. The public interest in protecting constitutional rights, ensuring democratic integrity, and maintaining the rule of law is paramount. These are not merely private interests of the Petitioners but fundamental public interests affecting all Kenyans. The protection of fundamental rights, the integrity of elections, and the principle that powerful technologies must be subject to democratic control and constitutional safeguards are core public interests that this Honourable Court is constitutionally obligated to vindicate. 4.4.2 Limited Impact on Legitimate AI Development The conservatory order sought restrains deployment of high-risk artificial intelligence systems. It does not prohibit all AI use, nor does it prohibit research, development, or low-risk applications of AI technology. High-risk AI systems are those that pose significant threats to fundamental rights for example, mass surveillance systems, biometric identification systems deployed in public spaces, automated decision-making systems in contexts like automated Instant Fines Traffic Management System, credit scoring or hiring without human oversight, AI systems used in law enforcement or criminal justice, and AI systems capable of generating deepfakes or manipulating electoral processes. The temporary restraint on deployment of such high-risk systems, while a comprehensive legal framework is being developed, represents a proportionate and reasonable limitation necessary to protect fundamental rights. It does not prevent the Respondents from developing a regulatory framework, conducting public consultation, or preparing for responsible AI adoption. 4.4.3 Respondents Can Act Swiftly to Develop Framework The substantive Petition seeks orders compelling the Respondents to develop and publish a draft National AI Policy Framework and to table draft AI governance Page 19 legislation within 90 days. This is a reasonable timeframe given the urgency of the matter and the fact that the Respondents already released a National AI Strategy in 2025. If the Respondents act with appropriate urgency, the conservatory orders need only be in place for a short period while a proper legal framework is developed. The inconvenience to the Respondents—being temporarily restrained from deploying highrisk AI systems without safeguards is minimal compared to the irreparable harm to constitutional rights and democratic processes that would result from allowing such deployment to continue. 4.4.4 International Best Practice Supports Precautionary Approach International human rights frameworks emphasize the need for precautionary approaches to AI governance. The African Commission on Human and Peoples' Rights, has called on African States to develop human rights-compatible AI governance frameworks. 21 KER188/2026 UNESCO's Recommendation on the Ethics of Artificial Intelligence emphasizes transparency, accountability, human rights impact assessments and effective remedies. Major jurisdictions including the European Union, United States, United Kingdom, Canada, Australia, Singapore, Japan and others have enacted or are enacting comprehensive AI governance frameworks that impose requirements for high-risk AI systems including mandatory risk assessments, transparency obligations, human oversight requirements and accountability mechanisms. Kenya should not be an outlier by allowing unregulated deployment of high-risk AI systems while other jurisdictions are establishing comprehensive safeguards. The public interest favours Kenya adopting international best practices and developing a framework that protects rights while enabling responsible innovation. 4.4.5 Risk of [Prejudice] If No Orders Granted If this Honourable Court declines to grant conservatory orders despite the Petitioners establishing a prima facie case and irreparable harm, it would set a precedent that highrisk technologies can be deployed without adequate safeguards so long as deployment happens quickly enough that litigation cannot prevent it. This would undermine the preventive protection that constitutional rights require and would render the Nubian Rights Forum principle meaningless. Conclusion on Balance of Convenience: The balance of convenience and public interest overwhelmingly favour the grant of conservatory orders. The public interest in constitutional rights, electoral integrity and rule of law far outweighs any inconvenience to Respondents from being temporarily restrained from deploying high-risk AI systems without adequate safeguards. 5. PRAYER In light of the foregoing submissions, the Petitioners respectfully pray that this Honourable Court be pleased to: GRANT the conservatory order restraining the 1st and 2nd Respondents, whether by themselves, their officers, servants, agents, or any person acting on their behalf, from deploying, authorizing and/or operationalizing the use of high-risk artificial intelligence systems, pending the hearing and determination of the Petition. 7. CONCLUSION The Petitioners have established a strong prima facie case that unregulated deployment of high-risk AI systems violates multiple fundamental rights guaranteed under the Constitution of Kenya 2010. The existing legal framework is inadequate, and the Respondents have failed to fulfil their positive constitutional obligations to establish safeguards before allowing deployment of technologies that pose systemic rights threats. Without conservatory orders, the Petitioners and all Kenyans will suffer irreparable harm to their constitutional rights and to democratic processes. The substratum of 22 KER188/2026 the Petition will be rendered nugatory if high-risk AI systems continue to proliferate during the pendency of this case. The balance of convenience and public interest overwhelmingly favour the grant of conservatory orders. The forthcoming August 2027 elections face unprecedented threats from AI-driven disinformation and manipulation. The integrity of Kenya's democratic processes cannot be secured if this litigation proceeds without interim protection. The principle established in Nubian Rights Forum & 2 Others v Attorney General & 6 Others, that comprehensive legal and regulatory frameworks must be in place before deployment of intrusive digital systems, applies with even greater force to AI systems that pose systemic threats to fundamental rights and freedoms. The Petitioners humbly pray that this Honourable Court will grant the conservatory orders sought.” [12] The Respondents by Submissions dated 1/4/2026 opposed the application as being based on speculation, as follows: “B. ISSUES FOR DETERMINATION 1. My lord it is our considered opinion that the following are the issues arising for your determination; Whether the Applicant has met the threshold for grant of conservatory orders. ANALYSIS 2. 3. 4. My we humbly submit that jurisprudence on the threshold for the grant of conservatory orders is extensive. In Wilson Kaberia Nkunja vs. The Magistrate and Judges Vetting Board and Others Nairobi High Court Constitutional Petition No.154 of 2016 (2016) eKLR, the court held that the applicant must; (a) An applicant must demonstrate that he has a prima facie case with a likelihood of success and that unless the court grants the conservatory order, there is a real danger that he will suffer prejudice as a result of the violation or threatened violation of the Constitution. (b) Whether, if a conservatory order is not granted, the Petition alleging violation of, or threat of violation of rights will be rendered nugatory: and (c) The public interest must be considered before grant of a conservatory order. On the issue of prima facie case My lord we submit that the Applicant has not placed before this Honorable Court any concrete evidence demonstrating violation or threatened violation of constitutional right arising from the conduct of the Respondents. The application is premised on the alleged absence of a comprehensive artificial intelligence regulatory framework and generalized assertions on the risks of Al systems. My lord a prima facie case must be grounded on identifiable facts. The applicants have failed to identify any specific artificial intelligence systems deployed, authorized or operationalized by the respondents that has violated or imminently threatens to violate their constitutional rights. The application relies on broad allegations relating to possible future harms such as electoral interference, discrimination, and economic disruption which remain speculative. 5. My lord we place reliance on the court's holding in Centre for Rights Education and Awareness and 7 others -v- The Attorney General [HCCP No. 16 of 2011]: 23 KER188/2026 "At this stage, a party seeking a conservatory order only requires to demonstrate that he has a prima facie case with a likelihood of success and that unless the court grants the conservatory order, there is real danger that he will suffer prejudice as a result of the violation or threatened violation of the Constitution". 6. Similarly, in Board of Management of Uhuru Secondary School v City County Director of Education & 2 others [2015] KЕНС 2174 (KLR) the court held that:. "It is in my view not enough to merely establish a prima facie case and show that it is potentially arguable. Potential arguability is not enough to justify a conservatory order but rather there must also be evident a likelihood of success. The prima facie case ought to be beyond a speculative basis." 7. Further in Martin Nyaga Wambora -v- Speaker of the County Assembly of Embu & 3 Others CP No. 7 of 2014 illustrated on the issue of real danger as follows: "To those erudite words I would only highlight the importance of demonstration of "real danger". The danger must be imminent and evident, true and actual and not fictitious; so much so that it deserves immediate remedial attention or redress by the court. Thus, an allegedly threatened violation that is remote and unlikely will not attract the court's attention." 8. The court in Martin Nyaga Wambora -v- Speaker of the County Assembly of Embu & 3 Others CP No. 7 of 2014, additionally stated the following in regards to the second principle: "flowing from the first two principles, is whether if an interim Conservatory order is not granted, the petition or its substratum will be rendered nugatory. It is indeed the business of the court to ensure and secure so far as possible that any transitional motions before the court do not render nugatory the ultimate end of justice." 9. We submit that the applicants have not demonstrated that the substratum of the petition is under any imminent threat. The petition concerns alleged absence or inadequacy of artificial intelligence regulatory framework which is a policy and legislative issue capable of adjudication at the hearing of the petition. There is no evidence that the respondents are undertaking any specific action that would defeat the petition or render it nugatory if the orders sought are not granted. 10. Lastly My lord we submit on the matter of public interest; we rely on the apex Court's decision in Gatirau Peter Munya v Dickson Mwenda Githinji & 2 Others SCK Petition No 2 of 2013, where Ojwang and Wanjala SCJJ. stated that: "Conservancy orders' bear a more decided public-law connotation: for these are orders to facilitate ordered functioning within the public agencies, as well as to uphold the adjudicatory authority of the court, in the public interest. Conservatory orders, therefore, are not, unlike interlocutory injunctions, linked to such private party issues as 'the prospects of 24 KER188/2026 irreparable harm' occurring during the pendency of a case; or 'high probability of success' in the supplicant's case for orders of stay. Conservatory orders, consequently, should be granted on the inherent merit of a case, bearing in mind the public interest, the constitutional values and the proportionate magnitudes, and priority levels attributable to the relevant causes, Thus, where a conservancy order is sought against a public agency like a legislative assembly that is mandated to carry out certain functions in the normal course of its business, it is only to be granted with due caution. The interruption of the lawful functions of the legislative body should take into account the need to allow for their ordered functioning in the public interest" 11. Similarly, in Wambora v Speaker of the County of Assembly of Embu & 3 others [2014] KEНС 7498 (KLR) the court cautioned that where the conservatory orders are sought against public bodies, the court must exercise restraint to not interfere with their lawful mandate unless exceptional circumstances are demonstrated. 12. My lord we wish to draw attention to this Honourable Court's holding in KEROGOYA HCCHRPET E001 of 2026 Green Thinking Action vs Hon Attorney General where this Honourable stated that: "conservatory orders should not, at an interlocutory stage, lightly interfere with fiscal or regulatory measures of general application absent compelling grounds". 13. My lord in the present application the orders sought seek to restrain the Respondents from deploying, authorizing and or operationalizing the use of high risk artificial intelligence. This prayer would have the effect of halting a wide range of government functions including this Honourable Court's without clarity or limitation. Granting such orders would unjustifiably interfere with the Respondent's constitutional and statutory mandate in the development and implementation of digital policy and services and would be contrary to public interest. 14. Accordingly, my lord we submit that Applicants have failed to satisfy the threshold for the grant of conservatory orders, and thus urge this Honourable Court to dismiss the application with costs.” [13] Although listed as co-petitioners the 2nd and 3rd Petitioners did not attend in person or by counsel to take part in the hearing of the application. [14] The 1st Petitioner filed a Supplementary Affidavit in purported response to the Ground of Opposition filed by the respondents herein long after the Respondents had filed their submissions on 1/4/2026. While the petitioner as applicant was entitled to grant of leave to respond to the response by the Respondent, the filing of the supplementary affidavit without leave after the Submissions and reservation of the ruling is improper. The Court, therefore, notes that the Respondent has not had opportunity to submit on the Supplementary Affidavit, and, consistently with the right to a fair hearing under Article 50(1) of the Constitution, it cannot therefore make it a basis for the grant or refusal of any 25 KER188/2026 orders without fair hearing granted to all parties. The Affidavit may, however, be used in further support of the Petition at the hearing thereof at the appropriate time. Amicus Curiae [15] The Amicus has filed Amicus Brief dated 23/3/2026 on the Petition itself. As regards the state of the evidence and particularisation of the alleged violations or specificity of the petition, as sometime called in terms of the leading decision of Anarita Karimi Njeru vs. Republic [1979] KEHC 30 (KLR), the Amicus submits as follows: “WHETHER THERE IS A VIOLATION OF FUNDAMENTAL RIGHTS AND FREEDOMS UNDER ARTICLES? 48. The amicus curiae submits that in determining whether or not there has been any contravention of the Constitution or the Petitioners’ fundamental rights and freedoms, the court should be guided by the principle that he who alleges must prove and that a party must plead with precisions and adduce supporting evidence whenever allegations pertaining to violations and contraventions or a threat to violations and contraventions of fundamental rights and freedoms and the Constitution in general have been made. 49. The foregoing principle is well founded in judicial pronouncements such as in the case of EG & 7 OTHERS VS ATTORNEY GENERAL; DKM & 9 OTHERS (INTERESTED PARTIES); KATIBA INSTITUTE & ANOTHER (AMICUS CURIAE), CONSTITUTIONAL PETITION NO. 150 OF 2016 where the court made the following observation thus: [303] The general principal governing determination of cases is that a party who makes a positive allegation bears the burden of proving it. Moreover, the onus to establish the violation of alleged rights is not a mere formality. Differently put, the onus lies on he who alleges to prove every element constituting his or her cause of action. This includes sufficient facts to justify a finding that the rights have been violated. [304] Constitutional analysis under the Bill of Rights takes place in two stages. First, the applicant is required to demonstrate his or her ability to exercise a fundamental right has been infringed. If the court finds that the law, measure, conduct or omission in question infringes the exercise of the fundamental right, or a right guaranteed in the Bill of Rights, the analysis may move to the second stage. In the second state, the party seeking to uphold the restriction or conduct will be required to demonstrate the infringement or conduct is justifiable in a modern democratic state and satisfies Article 24 test. [305] Cases are decided on the legal burden of proof being discharged (or not). Lord Brandon once remarked: 26 KER188/2026 “No Judge likes to decide cases on the burden of proof if he can legitimately avoid having to do so. There are cases, however, in which, owing to the unsatisfactory state of the evidence or otherwise, deciding on the burden of proof is the only just course to take.” [306] Whether one likes it or not, the legal burden of proof is consciously or unconsciously the acid test applied when coming to a decision in any particular case. This fact was succinctly put forth by Rajah JA in Britestone Pte Ltd vs Smith & Associates Far East Ltd : “The court’s decision in every case will depend on whether the party concerned has satisfied the particular burden and standard of proof imposed on him” [307] Decisions on violation of constitutional rights should not, and must not, be made in a factual vacuum. To attempt to do so would trivialize the constitution and inevitably result in ill-considered opinions. The presentation of clear evidence in support of violation of constitutional rights is not a mere technicality; rather, it is essential to a proper consideration of constitutional issues. Decisions on violation of constitutional rights cannot be based upon unsupported hypotheses.” “HIGH IMPACT ARTIFICIAL INTELLIGENCE” 50. Whereas the Petitioners have mentioned “HIGH IMPACT ARTIFICIAL INTELLIGENCE’ they have not gone further to elaborate on the same nor tender evidence on the deployment. In the absence of such an elaboration, we urge the Court to be guided by the principle in ANARITA KARIMI NJERU VS REPUBLIC (MISCELLANEOUS CRIMINAL APPLICATION 4 OF 1979) [1979] KEHC 30 (KLR) (Crim) (29 January 1979) (Judgment) which places an obligation on a person to plead with precision and specificity. The Court pointed out thus: “We would, however, again stress that if a person is seeking redress from the High Court on a matter which involves a reference to the Constitution, it is important (if only to ensure that justice is done to his case) that he should set out with a reasonable degree of precision that of which he complains, the provisions said to be infringed, and the manner in which they are alleged to be infringed.” 51. The foregoing notwithstanding, the Amicus invites the Court to consider various pronouncements on the different types and classifications of “Artificial Intelligence” systems as espoused by various authors, among others....” [16] The Amicus’ conclusion supports regulation by evolution through gradual legislation and regulation rather than immediate comprehensive regulation has been as follows: “WHETHER THERE HAS BEEN ALLEGED UNREASONABLE DELAY IN REGULATING ARTIFICIAL INTELLIGENCE? 27 KER188/2026 54. The Amicus Curiae submits that the allegation that the respondents have unreasonably delayed regulating artificial intelligence must be assessed in light of existing statutory obligations and any such comparative regulatory practice and the global evolution of Artificial Intelligence governance frameworks. 55. Comparative experience, as highlighted elsewhere herein, demonstrates that Artificial Intelligence regulation is being developed progressively across jurisdictions through phased and risk-based frameworks rather than immediate comprehensive legislation. A useful illustration is the European Union Artificial Intelligence Act, 2024 widely regarded as the first comprehensive legislative framework governing AI systems. 56. The European Union Artificial Intelligence Act itself recognises the complexity of regulating artificial intelligence and therefore adopts a tiered and phased regulatory model based on risk classification. Under this model, Artificial Intelligence systems are categorized into prohibited systems, high-risk systems, and minimal-risk systems, with regulatory obligations calibrated according to the level of risk posed to safety and fundamental rights. Importantly, the Act does not impose all obligations immediately. Instead, it establishes a progressive implementation timeline following its entry into force on 1 August 2024, with different obligations taking effect over several years. For instance, prohibitions on unacceptable AI systems take effect six months after entry into force, while most obligations applicable to high-risk Artificial Intelligence systems only become operative twenty-four months later, and some obligations relating to systems used by public authorities extend as far as 2030. 57. This phased approach reflects a recognition that Artificial Intelligence regulation requires time for institutional preparation, technical standard-setting, and regulatory capacity building. The European Union framework therefore emphasises the need for organisations first to identify and inventory Artificial Intelligence systems, assess their risk classification, and establish governance mechanisms before compliance obligations fully apply. 58. In addition, the European Union Artificial Intelligence Act focuses regulatory intervention primarily on high-risk systems capable of affecting fundamental rights, such as systems used in employment decisions, creditworthiness assessments, access to essential services, law enforcement, and democratic processes. 59. The Amicus Curiae submits that this comparative experience is instructive. Even in jurisdictions with advanced regulatory capacity such as the European Union, Artificial Intelligence governance has evolved through gradual legislative and regulatory development rather than immediate comprehensive regulation. 60. The Amicus Curiae’s view is that the assertion of unreasonable delay must be evaluated in light of the Government’s ongoing policy initiatives, including the mentioned Strategy, which outlines a roadmap for developing governance frameworks, ethical safeguards, and regulatory oversight mechanisms for Artificial Intelligence technologies.” 28 KER188/2026 [17] The Amicus, however, supports the judicial intervention for the grant of the Petition of appropriate relief to enforce enactment of the necessary legislation to protect violations of the rights and fundamental freedoms, citing historical experience with related matters. Determination [18] The principles upon which the Court considers an application for conservatory orders are well known and settled by various caselaw authorities of the High Court, the Court of Appeal and ultimately the Supreme Court of Kenya. See for example the decision cited by the parties here, Gatirau Peter Munya v Dickson Mwenda Githinji & 2 Others SCK Petition No 2 of 2013; Centre for Rights Education and Awareness (CREAW) & 7 Others –v- Attorney General [2011] eKLR; and Board of Management of Uhuru Secondary School v. City County Director of Education & 2 Others [2015] eKLR. [19] This Court has also had its share of consideration of teh principles for grant of conservatory orders as shown in as Nairobi Const. Pet. NO. 206 OF 2016 SATINDERJIT SINGH MATHARU v. ARMAJIT SINGH GAHIR & 5 OTHERS, where the Court held: “Principles for the grant of the conservatory order 5. Despite varied nomenclatural expressions, the principles upon which the High Court considers application for conservatory orders in constitutional litigation are now settled by several decisions on the point, and may be condensed as follows:- 1. The applicant must demonstrate prima facie case, or an arguable case, for the grant of the relief sought. 2. The applicant must stand to suffer an irreparable harm, injury or loss not remediable by any other relief; and 3. As a remedy in constitutional litigation, the conservatory order calls for consideration of the public interest in the matter, and the balance of convenience between the petitioner’s and the respondent’s case must favour the grant of the conservatory order. See generally, authorities cited by the petitioner, AG v. Sumair Bansraj [1985] 38 WIR 286; Tunoi & Anor v. Judicial Service Commission and Anor, (2014) eKLR; Judicial Service Commission v. Speaker of the National Assembly & Anor. (2013) eKLR; Gatirau Peter Munya v. Dickson Mwenda Kithinji & 2 Ors. (2014) eKLR; Centre for Rights Education and Awareness (CREAW) & 7 Ors. v. AG (2011) eKLR; and Centre for Human Rights and Democracy & Ors. v. The Judges and Magistrates Vetting Board (2012) eKLR. 6. This Court has also previously expressed itself on the matter. See Muslims for Human Rights (MUHURI) & 4 Ors. v. Inspector General of Police & 2 Ors., Mombasa PETITION NO. 62 OF 2014 of 22nd December 2014, where the Court held as follows: “The emerging principles for the grant of injunction or conservatory orders under the constitutional litigation, as I 29 KER188/2026 understand them, are firstly, that the applicant must demonstrate an arguable case - sometimes called prima facie arguable case - the reference to arguable case distinguishing it from the prima facie test of the Giella v. Casman Brown (1973) EA 385 traditionally applied in regular civil cases; secondly, that the applicant must show that the petition would be rendered nugatory or that the damage that would be suffered in the absence of the conservatory order would be irreversible; and, thirdly, that in constitutional cases, the public interest in the matter would be considered and generally upheld. See Kenya Transport Association Limited v. Cabinet Secretary for Transport and Infrastructure and Ors., Mombasa HC Petition No. 16 of 2014 where I considered some of the decisions on the matter as follows: “The tests for the grant of conservatory orders has been variously expressed by different courts. See Mombasa High Court petition No. 7 of 2011, Muslim for Human Rights and 2 Ors v the Attorney General, per Ibrahim J. (as he then was), Mombasa High Court Petition No. 47 of 2011 Harun Barky Yator v. Judicial Service Commission (JSC), per Okwengu J, (as she then was), Nairobi High Court Petition No. 557 of 2013, per Majanja J, and Mecha Magaga v Jackson Obiero Magaga (2014) eKLR, per Okong’o, J. All the courts require for the grant of conservatory orders a prima facie case or a prima facie arguable case as in Yator’s case; irretrievability or irreparability if conservatory order is not granted and the subject matter is irretrievably lost (akin to the irreparability of damage test) and a balancing of the interests of the applicant and the respondents. There arises confusion as to whether the test of standard of the applicant’s case is on the prima facie or arguable case. Once accept that the court cannot determine the disputed merits of the case at the interlocutory stage, the correct standard must be the standard of arguable case. See Mbuthia v. Jimba Credit Corporation (1988) KLR 1. I also consider that Under Article 23 (3) of the Constitution, the court may make a broad spectrum of orders as conservatory orders to preserve the status quo where circumstances warrant and that may include 30 KER188/2026 fashioning a remedy to fit the particular circumstances of the application before the court.’” [20] The conservatory order test as rendered in differing terminology by the various courts require an applicant for conservatory order to demonstrate – a. arguability of his case by a prima facie case for the grant of the reliefs sought; b. Prejudice or irreparable harm if conservatory order is not granted; that he will suffer substantial loss or the Petition shall be rendered nugatory; and c. Public interest consideration in view of the public nature of the proceedings, as a balance of convenience, favours the grant of the conservatory order. [21] The Petitioners have set out the possible violations of various rights that may be violated by deployment of AI systems in various areas of governance in the Country. An illustration of such AI systems is given in the Petition and supported by a submission that “artificial intelligence systems are being deployed across Kenya in contexts affecting fundamental rights including government services, financial services, employment, education, law enforcement, healthcare, and digital platforms without adequate legal safeguards, transparency requirements, accountability mechanisms, or effective remedies”. There is, however, no evidence in the supporting affidavit as to identity of such AI systems and where they are deployed with the result of the alleged effect of violation or threats to violation of rights. [22] The best case for the Petitioners’ to the prayer for conservatory orders is set out in the Submissions as follows: “The conservatory order sought restrains deployment of high-risk artificial intelligence systems. It does not prohibit all AI use, nor does it prohibit research, development, or low-risk applications of AI technology. High-risk AI systems are those that pose significant threats to fundamental rights for example, mass surveillance systems, biometric identification systems deployed in public spaces, automated decision-making systems in contexts like automated Instant Fines Traffic Management System, credit scoring or hiring without human oversight, AI systems used in law enforcement or criminal justice, and AI systems capable of generating deepfakes or manipulating electoral processes. The temporary restraint on deployment of such high-risk systems, while a comprehensive legal framework is being developed, represents a proportionate and reasonable limitation necessary to protect fundamental rights. It does not prevent the Respondents from developing a regulatory framework, conducting public consultation, or preparing for responsible AI adoption. ” [23] The only shortcoming in the case is that the evidence as to which these AI technologies, where they are deployed and their capabilities and propensity to violate rights is not disclosed in he Petition. Mere illustration of the AI systems as “mass surveillance systems, biometric identification systems deployed in public spaces, automated decision-making systems in contexts like automated Instant Fines Traffic Management System, credit scoring or hiring without human oversight, AI systems used in law enforcement or criminal justice, and AI systems capable of generating deepfakes or 31 KER188/2026 manipulating electoral processes” is not necessary particulars and evidence of deployment by the Respondents. [24] What AI systems by name, area and place of deployment is the Court being asked to halt? Court orders must not be vague. The exact act sought to be restrained must be identifiable with exactitude. Contempt of Court applications may follow and it should be no answer that the order of the Court was vague and ambiguous. Most importantly, in affording a fair hearing in accordance with the Article 50 (1) of the Constitution, the defendant/respondent must know the exact nature of the case he has to meet and respond to, and the Court must know the exact dispute that its is called upon to determine. For this to happen, clear and particularised pleadings and evidence in support must be availed by the plaintiff/petitioner to enable the defendant/respondent know the case he is called upon to answer. [25] While the petitioner’s case was strong on case-law applicable to the matters in issue, with which there was not much disagreement from the Respondents, there is only a theoretical submission that “that unregulated deployment of violates multiple fundamental rights high-risk AI systems guaranteed under the Constitution of Kenya 2010.” The high-risk AI systems, have not been identified by their types, areas of deployment and uses as to link them with alleged violations or threatened violations of the rights protected under the Bill of Rights of the Constitution. [26] Vague references to high-risk AI systems in certain areas of the state functions do not identify the impugned AI systems the subject of the application for conservatory orders. Only on a technical point, the conservatory order sought in the Notice of Motion is not supported by a substantive relief for injunctive relief. However, on consideration of the authority to fashion appropriate relief for breach or threatened breaches of the Bill of Rights, the Court does not hold this default against the petitioner, and in any event, it may be cured by amendment as necessary. [27] There is, however, as noted by the Respondents and the Amicus a paucity of evidence clearly linking any deployment of identifiable types of AI technologies to alleged breaches of the rights of persons as alleged. The general provisions of the law of evidence in sections 107-109 of the evidence Act and the particular case-law stipulation for particularity of constitutional claims as settled by the specie of cases following Anerita Karimi Njeru, place the factual evidential burden on the petitioners to put before the Court a warrant for the grant of the conservatory orders sought and eventually the substantial reliefs sought in the Petition. [28] The Court notes that the Petitioners’ claim as to violation, or threatened violation, of rights enumerated in the Petition allegedly by deployment and use by the Respondents of AI technologies is not substantiated by evidence of such deployment and use of AI systems which have been identified by their types, area and manner of deployment. There is room for better particularisation to enable the respondents to respond to the particular allegations and the Court to identify, demarcate and determine the dispute it is called to adjudicate. [29] In the absence of demonstration of a cause-and-effect relationship between a particularised AI system , which is shown to have been deployed and used in a particular area of governance process as alleged and the violation or threatened violation of a particular right alleged to be violated or threatened 32 KER188/2026 with violation, the Petitioner cannot demonstrate a prima facie case to a petition for constitutional reliefs against violation of rights and fundamental freedoms. [30] The Conservatory Order bearing, as held in Gatirau Munya case, a more public interest nature than the private law remedy of injunction must be granted only where the interests of the public coincide with the prayers in the Petition being granted immediately so as not to render the Petition nugatory. In this case, it is difficult to ascertain the public interest in the matter which is not sufficiently identified with particulars to enable examination of the effect of the alleged deployment of AI in the particular field to determine its potentiality to adversely affect rights and fundamental freedoms as alleged. It is not possible to say that the Petition will be rendered nugatory if the AI systems are employed without legal regulation, when those particular AI systems are not identified and their deployment and use and their effect or likely effect pleaded as to permit the Respondents to respond in a fair hearing leading to a judgment on the merits and an appropriate interlocutory remedy, as necessary. [31] A prima facie case in this case should leave the Court with no question as to the particular AI system, the deployment and use of the AI system and the resultant adverse effect, or likely adverse result, on a person’s or persons’ rights and fundamental freedoms The caselaw authority of Nubian Rights Forum & 2 Others v Attorney General & 6 Others [2020] eKLR (Huduma Namba case), cited by the Petitioner for proposition that “that comprehensive regulatory framework and adequate data protection framework must be in place before deployment of intrusive digital systems involving mass collection and processing of biometric and personal data, and that proceeding without such safeguards violates fundamental rights and constitutional principles [and] the principle applies with even greater force to artificial intelligence systems, which by their nature involve automated processing, profiling, pattern recognition and decision-making capabilities that can have far-reaching and systemic impacts on fundamental rights, democracy and the rule of law”, would strongly apply here, if such allegedly offending AI systems, like the Huduma Numba system in that case were identified. [32] The Petitioner has attempted hastily to cobble a listing of AI systems allegedly deployed and used to deleterious or likely detrimental effect in the illustration set out in Submissions filed on the Notice of Motion, as shown above. However, submissions are not pleadings or evidence in the matter and a determination on the merit of the Petition may only be made on ‘evidence’ put before the Court by way of submissions. In any event these allegedly high risk AI systems are not identified with necessary specificity. For a proper inquiry in as serious a matter as the issue before the Court cogent evidence is necessary. [33] With no pretensions to activist stance in the matter, the Court considers it in the public interest to facilitate full argument of the dispute on the merits; and for that purpose, pursuant the overriding objective of The Constitution of Kenya (Protection Of Rights And Fundamental Freedoms) Practice And Procedure Rules, 2013 (the Mutunga Rules) under Rule 3(2) thereof “to facilitate access to justice for all persons as required under Article 48 of the Constitution ”, the Court of its own motion shall permit the Petitioner to amend his petition to particularise the complaint, which the Court finds the public to have an interest in its effective and expedited determination. 33 KER188/2026 [34] The Supreme Court of Kenya in Mitubell case, [Mitu-Bell Welfare Society v Kenya Airports Authority & 2 others; Initiative for Strategic Litigation in Africa (Amicus Curiae) (Petition 3 of 2018) [2021] KESC 34 (KLR)], has approved the remedy of structural interdict as appropriate relief in constitutional litigation, and this Court considers the nature of the dispute herein on the enactment of legislation to regulate AI adoption and use in the country to require the Respondents as the responsible government agencies to report to court for purposes of informed further deliberation on the matter, the situational report on process and progress towards, and the existing measures, for the prevention and amelioration of any violations or threatened violations of the rights and fundamental freedoms of the persons. Conclusion [35] The Court is persuaded of the urgency of the matter in view of the rapid developments in the AI systems in the everyday life processes which impact on the lives of the people in employment, education, health and financial sectors and in particular the forth-coming 2027 elections as pointed out by the petitioners. [36] The Court is persuaded of the likely dangers of violations or threats of violations to related rights and fundamental freedoms of privacy, dignity, discrimination, property and consumer rights among others as presented by the Petitioners, if evidence of such deployment was adduced. [37] The Petition, as drawn presently, does not give specific particulars of the offending or likely offending AI systems and where they are deployed to enable targeted reply by the respondents and consideration by the Court for appropriate judicial intervention. Consequently, as such necessary evidence of such deployments has not been produced, it is not possible to specifically point to these offending systems so as to fashion an interim remedy halting their use until the necessary protective legislation and policy directions are put in place to protect irreparable and irreversible harm on our people in the meantime. [38] The Court does not, therefore, give a nonspecific ambiguous cover by a generic conservatory order against so called “high-risk AI systems”. The offending AI systems must be identified so that targeted conservatory order or remedial measure is granted. The Petition should be heard on the merits to identify areas of possible threats, if any, and make judicial intervention as necessary to give appropriate relief by way of guardrails against possible breach and provide a remedy of violations, past, present or future. [39] The Court considers, however, that there is a case for regulation of the AI technologies, and the Amicus brief supports such a case, with consideration being only whether immediate intervention by judicial process for coercive orders or a gradual growth of regulation policy and legislation is to be adopted. Consequently, the Court does not finds that the Petition should be struck out for want of particularity on the test in Anerita Karimi Njeru. [40] The Court rather feels compelled by the access to justice principle constitutional litigation in terms of Article 22 (3) (b) of the Constitution that a petition for redress of violations or threatened violation of human rights, consistently with the epistolary jurisdiction, may be moved by informal documentation, to grant the opportunity for the Petitioners to amend the Petition to give necessary 34 KER188/2026 particulars and support it with appropriate evidence of the deployment, use and likely infringement of rights and fundamental freedoms. [41] The Court does not consider that it is desirable or appropriate to hamper growth of AI systems; it is their adverse effect on the life, rights and fundamental freedoms of the people that must be safeguarded; and to do so, cogent evidence of deployment of AI technologies in government processes, as alleged, is necessary. That is the onus of the Petitioners. [42] A vague moratorium on the deployment and use of AI systems would, therefore, appear unwarranted. There is need for focus on the identified systems which cause or are likely to cause adverse effects on the people’s lives with regard to legal, social, economic, political, constitutional and human rights entitlements. Consequently, at this stage of hearing, the Court will not grant any conservatory orders on the basis of the Petition/application as drawn, for want of particularisation of the impugned AI systems. [43] However, in public interest and accepting that the problem, if any, does not go away, if the Court closes it eyes or buries its head in the sand, the Court will allow appropriate and necessary amendment to be effected by the Petitioner to bring into focus, response and consideration the offending AI systems that may require urgent of urgent and rapid intervention by a conservatory order of the Court. Appropriate relief [44] The Court has noted from the publication exhibited by the Petitioner of the Kenya AI Strategy 2025-2030 dated March 2025, the Kenya Government’s phased strategy to AI development and regulation as follows: “The strategy proposes a phased approach to ensure effective implementation. This begins with foundational investments in policy, infrastructure, and capacity building, followed by key milestones such as developing a national AI policy, establishing AI research and innovation hubs, executing pilot projects, and creating a monitoring and evaluation framework to track progress.” [45] The Court has also noted from the evidence so far presented before Court that the Petitioner had petitioned Parliament in July 2025 to enact legislation of regulation of Artificial Intelligence to promote ethical use of, protect fundamental rights and foster innovation and there is in existence a Bill pending debate and enactment in the National Assembly. It is important that as the Court considers making any or further orders/directions at the hearing of the Petition, as necessary, it should be served with information of the status of AI deployment and use and the existing regulatory frameworks so far and any policy and strategy for, and progress towards, a national general legislation therefor. [46] The Court further considers that the matter of protection of violation of human rights is the constitutional province of the Kenya National Humna Rights Commission (KNCHR) in terms of its mandate under Article 59 of the Constitution. The private petition before the Court might so much more be enriched by participation of the constitutional Commission charged with protection of human rights in all respects. The Court is mindful of its authority for joinder of parties under Rule 5 (d) (ii) of the Mutunga Rules that “the Court may at any stage of the proceedings, either upon or without the 35 KER188/2026 application of either party, and on such terms as may appear just— (ii) that the name of any person who ought to have been joined, or whose presence before the court may be necessary in order to enable the court adjudicate upon and settle the matter, be added” and under Rule 7(2) that “(2) A court may on its own motion join any interested party to the proceedings before it.” [47] The Court considers that the Kenya National Commission on Human Rights (KNCHR) is a necessary/Interested Party in this Petition in view of its constitutional functions especially under Article 59 (2) (c) and (d) “(c) to promote the protection, and observance of human rights in public and private institutions; [and] (d) to monitor, investigate and report on the observance of human rights in all spheres of life in the Republic, including observance by the national security organs ”. The Court shall, therefore, of its own motion direct that the KNCHR be joined as a necessary and Interested Party, with liberty to file a response to the Petition within the same time lines as allocated to the Respondents and Interested Party herein. [48] Consequently, while declining the conservatory order sought herein to stop the deployment of unnamed AI systems, the Court considers that sufficient evidence has been adduced to warrant the fashioning of appropriate remedy by way of structural interdict, (a constitutional remedy that the Supreme Court in Mitubell has validated in Kenya), to require the Respondents by their relevant bodies to file, with the Court with copies to the parties, a report on any action and progress so far towards the eventual regulation of deployments and use of AI technologies in Kenya. [49] There shall be liberty to apply. ORDERS [50] Accordingly, for the reasons set out above, the Court finds merit in the application and it is allowed to the extent and in terms as follows: 1. The prayer no. (iii) of the Notice of Motion for an order for priority hearing is granted and the Petition shall be admitted for hearing on priority basis and set for hearing on day-to-day basis on the dates 16-18 June 2026, and Judgment is reserved for Monday 29 th June 2026. 2. The prayer No. (ii) for a conservatory order restraining deployment of Artificial Intelligence (AI) systems in the “high-risk artificial intelligence systems” is vague and cannot be issued in the terms as prayed. 3. In the interests of justice leave is granted to the petitioner to amend as necessary to plead the particulars of the AI system that he considers “high- risk artificial intelligence systems” and which are the targets of his petition to enable the Respondents to specifically respond, and the Court to determine the nature and extent of the relief sought. 4. The Kenya National Commission on Human Rights (KNCHR) shall be joined to the Petition as a necessary/Interested Party and served with the amended Petition with liberty to file a response to the Petition within the same time lines as allocated to the Respondents and Interested Party herein. 36 KER188/2026 5. The necessary amendment to the Petition shall be effected and served within the next fourteen (14) days, that is to say by the 15 th May 2026 and the Respondents shall file their response to the amended petition within fourteen (14) days of service not later than 29 th May 2026 . Thereafter the Petitioner shall have seven (7) days that is by the 8 th June 2026 to file a supplementary affidavit in reply, if necessary. 6. The Respondents by their relevant bodies shall file with the Court with copies to the parties, situation report(s) on any action and progress so far towards the eventual regulation of deployments and use of AI technologies in Kenya within the next thirty (30) days, that is by the 29/5/2026. 7. There shall be liberty to apply for any interlocutory order/directions before then, as necessary. 8. Costs in the Cause. Order accordingly. DATED AND DELIVERED THIS 30TH DAY OF APRIL 2026. EDWARD M. MURIITHI JUDGE Appearances: Mr. John Wangai the 1st Petitioner. N/A for the 2nd & 3rd Petitioners. Mr. Kiongo with Ms. Nyakora for the 1st, 2nd, & 3rd Respondents. Mr. Sore for the Interested Party. Mr. Ochieng Odinga for the Amicus Curiae. 37