The Judiciary of Kenya, in collaboration with Strathmore University's Centre for Intellectual Property and Information Technology Law, has published Draft 2.0 of an Artificial Intelligence Policy for the Judiciary, the first formal governance framework for AI use across all levels of Kenya's court system. The Policy applies to every actor within the Judiciary, from the Chief Justice and judges to registrars, support staff, ICT professionals, and third-party technology vendors. Its publication marks a significant step in Kenya's engagement with one of the most consequential questions facing legal systems worldwide: how should courts govern the use of AI without undermining the constitutional values, independence, fairness, transparency, and access to justice, on which judicial authority depends?
The Policy operates within Kenya's existing legal framework, drawing its constitutional anchors from Articles 10, 31, 47, 48, 50, 159, and 160 of the Constitution of Kenya, 2010. It also operates alongside the Data Protection Act, 2019, the Judicial Service Act, 2011, and the Computer Misuse and Cybercrimes Act, 2018. Rather than positioning itself as a technical document, the Policy frames AI governance as a fundamentally constitutional exercise: the question is not merely what AI can do in courts, but whether it can do so in a manner consistent with Kenya's constitutional commitments. This framing, rooting technology governance in constitutional obligation rather than institutional preference, is among the Policy's most commendable features.
The Policy establishes nine foundational principles that govern AI adoption across the entire lifecycle of any AI system deployed within the Judiciary, from initial procurement through to decommissioning. These principles include Legality and Trustworthiness, Inclusivity and Accessibility, Fairness and Bias Prevention, Transparency and Explainability, Human Oversight and Accountability, Security and Reliability, Judicial Independence, Efficiency and Sustainability, and a distinctly Kenyan principle described as Locally Led, which requires that AI tools reflect the communities served by the Judiciary, prioritise locally developed or locally validated innovations, and ensure that data is stored and managed within Kenya where appropriate. Crucially, the Policy states that these principles are non-negotiable: any AI initiative that cannot demonstrate alignment with them is to be prohibited.
On governance, the Policy establishes a multi-tiered oversight structure anchored in an AI Governance Committee (AIGC), which is to be constituted from the existing Integrated Case Management Systems (ICMS) Committee. No AI system may be deployed within the Judiciary without the AIGC's explicit approval. The AIGC is responsible for approving AI-related projects, monitoring compliance, and advising the Chief Justice on emerging AI risks. Roles are clearly assigned across the Chief Justice and Judicial Service Commission, the AIGC, judges and judicial officers, the ICT Directorate, the Kenya Judiciary Academy, and third-party vendors. The structure is sound in design, but the Policy's requirement of mandatory AIGC approval for every deployment, without any de minimis exception or risk-tiered approval threshold, creates a potential bottleneck. A registry clerk wishing to use a transcription tool and a court deploying an AI-assisted sentencing system are treated identically under the current drafting, which risks making the approval mechanism either a chokepoint or a formality.