Kenya Artificial Intelligence Bill 2026: What the Proposed AI Regulator Would Oversee
By Shiled Jerono | October 1, 2026
Kenya’s Senate is considering an Artificial Intelligence Bill that would create a dedicated regulatory framework for the development and use of AI systems.
A Proposed Artificial Intelligence Commissioner
The Bill proposes an Artificial Intelligence Commissioner with responsibilities that would include oversight, compliance and development of ethical guidance.
Rules for High-Risk AI
High-risk AI providers would face additional obligations. The Bill proposes annual compliance reporting and transparency around the nature, purpose and limitations of automated systems.
Where automated decisions produce significant legal or similar effects, the Bill links safeguards to Kenya’s Data Protection Act, including human intervention and an opportunity to contest decisions.
Regulatory Sandboxes
The proposed regulator would also establish controlled regulatory sandboxes where AI systems could be tested under defined conditions before wider deployment.
Why the Bill Matters
The proposal comes as artificial intelligence is increasingly used in finance, recruitment, customer service, education and public administration. The policy question for lawmakers is how to encourage innovation while creating safeguards around bias, privacy and accountability.
The Bill Is Still a Proposal
The provisions may change during parliamentary consideration. They should not be treated as final regulatory obligations until the legislative process is complete.
What Counts as High-Risk AI?
The Bill defines a high-risk artificial intelligence system as one that may create significant risks to health, safety, fundamental rights or wider societal welfare. That category could cover systems used in sensitive decisions such as employment, access to services, finance or other areas where an automated decision can materially affect a person.
Transparency and Human Review
The proposal would require greater transparency from providers and users of high-risk systems. It also connects significant automated decisions to safeguards already recognised under data-protection law, including the possibility of human intervention and a chance to challenge an automated outcome.
How the Bill Balances Regulation and Innovation
The regulatory-sandbox model is intended to let developers test products in a controlled environment before wider deployment. That approach can allow regulators to observe emerging risks without automatically blocking experimentation.
For Kenyan businesses, the final compliance burden will depend on which systems are eventually classified as high risk, what reporting rules are prescribed in regulations and how the proposed AI Commissioner’s office coordinates with existing regulators such as the Office of the Data Protection Commissioner.
