A Senate committee has recommended that the Senate not proceed with the Artificial Intelligence Bill, 2026, sponsored by nominated Senator Karen Nyamu, The Star reported on 15 October.
The Information, Communication and Technology Committee, chaired by Trans Nzoia Senator Allan Chesang, said it considered submissions from 143 stakeholders. It warned that the Bill "may regulate AI too broadly, too early and too punitively for Kenya's current stage of digital and innovation ecosystem development".
The Bill, introduced in the Senate in April, would create an Office of the Artificial Intelligence Commissioner with powers to register AI systems, investigate complaints, conduct audits and enforce compliance, according to The Star and Techweez. It also proposed an AI Authority to set standards and run testing sandboxes, and an Advisory Committee. Serious offences would carry fines of up to Sh5 million, imprisonment for up to two years, or both.
The committee said those penalties could disproportionately punish small developers whose work is experimental and non-commercial. It also objected to what stakeholders called a "Super Regulator", warning that the Commissioner's office could duplicate the work of the Office of the Data Protection Commissioner and the Communications Authority of Kenya, and raised the cost of creating new institutions. It faulted the Bill for not adequately addressing cross-border AI systems and general-purpose AI models.
Microsoft Kenya proposed placing AI regulatory functions within existing institutions, The Star reported. Google and Safaricom raised concerns about the Bill's definitions and obligations.
The committee recommended that the government first finalise a national policy. The Ministry of ICT told it that the National Artificial Intelligence and Emerging Technologies Policy, 2026, was in its final stages.
Techweez notes that the committee rejected the Bill's approach, not the idea of regulating AI, and that Kenya still has no dedicated AI law. It says the Data Protection Act, 2019, covers some ground, including automated decision-making, but was not designed for questions such as algorithmic accountability or non-consensual deepfakes. The reports do not say when the full Senate will consider the committee's recommendation.