Kenya has prepared a landmark legislation that will create a powerful artificial intelligence (AI) watchdog to ensure ethical use of the technology and protect privacy and jobs.
The Kenya Artificial Intelligence Bill, 2026, introduces the Office of the Artificial Intelligence Commissioner to review, approve, flag risks and monitor AI systems in companies and government bodies.
The sector has seen explosive growth across the world -- driving huge profits but also stoking fears about bias, privacy and even the future of humanity.
The Bill is proposing the classification of AI products according to risk and adjusting scrutiny accordingly.
It also places Kenya in the mix of global attempts to address the dangers associated with AI.
“To provide a comprehensive framework for the regulation and governance of artificial intelligence in Kenya, ensuring ethical, transparent, and accountable use while fostering innovation and safeguarding human rights, data protection, and public welfare,” says the Bill.
The main idea of the proposed law is to regulate AI based on its capacity to cause harm to society. The higher the risk, the stricter the rules.
Kenya seeks to classify AI systems based on the level of risk they pose to society in the quest to create a trustworthy artificial intelligence.
The category of risks includes unacceptable risk, high risk, limited risk and minimal risk, with technologies considered to pose “unacceptable risk” banned outright.
The draft law does not explicitly give examples of such systems, but it borrows from the European AI Act.
In Europe, AI applications that pose a “clear risk to fundamental rights” face bans.
Examples include those that involve the processing of biometric data.
AI systems considered “high-risk”, such as those used in critical infrastructure, education, healthcare, law enforcement, border management or elections, will have to comply with strict requirements.
Low-risk services, such as spam filters, will face the lightest regulation, borrowing heavily from the EU, which expects most services to fall into this category.
Experts have also raised an alarm about the unconsented use of personal data to train models, unauthorised profiling, misinformation, discrimination against certain demographics, and a lack of accountability for mistakes made by AI tools.
“A person who designs or deploys an artificial intelligence system shall design or deploy the system in a manner that enhances rather than replaces human capabilities,” says the proposed law.
“Incorporate features that support human involvement in the system; and provide for human oversight in critical decisions made by the system.”
For AI systems likely to affect jobs, firms would be required to conduct workforce impact assessments evaluating potential job displacement and introduce mitigation measures such as reskilling programmes.
The Bill proposes fines of up to Sh5 million or imprisonment for up to two years for offences such as deploying prohibited AI systems, failing to conduct required risk or workforce assessments, or distributing harmful AI-generated content using someone’s likeness, commonly known as deepfakes, without consent.
Executives could also individually face liability for offences committed by their companies.
None of Africa’s 55 countries has developed a dedicated law regulating AI systems; 16, including Kenya, only have national AI strategies.
The proposed law establishes regulatory sandboxes where companies can test and pilot AI systems under regulatory supervision.
The AI commissioner would set conditions for participation in the sandboxes, prioritising innovations aligned with “national priorities” and collaboration with county governments.
“A provider of a high-risk artificial intelligence system shall submit annual compliance reports to the Artificial Intelligence Commissioner, and non-confidential information from the reports shall be made available to the public,” the Bill says.
Kenya’s draft law mirrors the European Union Artificial Intelligence Act, the world’s first comprehensive AI law, which came into force in August 2024.
The EU law also requires that AI tools remain under human oversight rather than being left entirely to automated decision-making to prevent harmful outcomes.
But unlike the Kenya draft law, the European rules explicitly prohibit applications considered to pose “unacceptable risk”, including social scoring systems that classify people based on behaviour or socio-economic characteristics, and certain forms of biometric identification.
