Kenya Senate Committee Rejects AI Bill

Kenya's Senate committee responsible for information and communications technology has recommended that the chamber should not proceed with the Artificial Intelligence Bill, 2026, arguing that the proposal could regulate a young technology sector too broadly and too punitively.

The recommendation is important, but it is not the same as a final vote by the full Senate. The committee examined the bill and submissions from 143 stakeholders before concluding that Kenya should complete a national AI policy and reconsider the design of legislation rather than enact the proposal in its present form.

The bill, sponsored by nominated Senator Karen Nyamu, proposed an Office of the Artificial Intelligence Commissioner with powers to register systems, investigate violations, conduct audits and impose sanctions. Serious offences could attract fines of up to KES5 million, imprisonment of up to two years, or both.

The Star reported that the committee was concerned about harsh penalties for small developers, the cost of creating another regulator and overlap with existing institutions such as the Office of the Data Protection Commissioner and the Communications Authority of Kenya. Microsoft Kenya reportedly argued that AI oversight could be placed within existing bodies, while Google and Safaricom raised concerns about definitions and obligations.

The committee's official report provides the institutional basis for the recommendation. Its rejection of this bill should not be read as a rejection of AI governance. Kenya already has a national AI strategy, and the ICT ministry told lawmakers that a National Artificial Intelligence and Emerging Technologies Policy was nearing completion.

The real disagreement is about sequencing and regulatory design. A broad law can become difficult to enforce when most advanced models are developed abroad and supplied through international platforms. Creating a powerful commissioner does not automatically give Kenya technical access to those systems or the leverage to audit them effectively.

At the same time, delaying legislation carries risks. AI is already being explored in Kenyan public services, elections, finance, education and correctional institutions. Citizens need enforceable rights when automated systems use personal data, affect access to services or create harmful synthetic media.

A revised approach could begin with risk rather than a blanket registration requirement. Systems used for employment, credit, healthcare, policing or essential public services deserve stronger documentation, testing and appeal mechanisms than low-risk productivity tools. Existing regulators could receive additional technical capacity while a central coordinating body develops common standards.

Kenya should also preserve room for supervised experimentation. Regulatory sandboxes can help local developers test systems under clear safeguards, but they should not become exemptions from data protection or consumer rights. Smaller companies need proportionate compliance, published guidance and access to testing resources rather than uncertainty backed by criminal penalties.

The committee recommendation gives Kenya an opportunity to improve the law before it hardens into an expensive structure. The next version should be built through transparent consultation and should explain which harms it addresses, which institution is responsible and how citizens can obtain a remedy.

For now, the Artificial Intelligence Bill has encountered a decisive institutional objection, not the end of Kenya's regulatory debate. The quality of what replaces it will show whether the country can protect innovation and public rights without treating them as opposing goals.

Sources: Kenya Parliament and The Star.