Governing the Machine: Power and Responsibility in Kenya’s AI Bill 2026

The Artificial Intelligence Bill, 2026, arrives at a moment when societies are grappling with questions of governing intelligence that is no longer exclusively human. Beneath its formal language and institutional design, the Bill reads as an attempt to reconcile competing visions of the future—one driven by innovation and possibility, the other by caution and control. It is, at its core, a document about power: who holds it, how it is exercised, and how it is restrained in a world where decision-making is increasingly shaped by machines.

In seeking to foster innovation while constructing a regulatory framework, the Bill embodies a familiar tension within modern governance: establishing an Artificial Intelligence Commissioner and introducing systems of classification, audit, and enforcement reflect a belief that technological progress must remain aligned with the public interest. Yet, innovation rarely thrives under rigid constraint, and the attempt to both enable and limit AI reveals a deeper puzzle: progress depends on freedom, but freedom without structure risks harm—meaning the Bill does not resolve this contradiction so much as it formalises it.

This contradiction carries through into the Bill’s emphasis on “human-centric” artificial intelligence, where the insistence on human oversight and accountability appears reassuring but is complicated by the growing complexity of AI systems, which risks reducing oversight to a procedural formality.

It is perhaps in recognition of these limits that the Bill introduces regulatory sandboxes, allowing experimentation within controlled environments and signalling that governance must evolve alongside the technology it seeks to regulate. What emerges, then, is not a resolution but a careful structuring of competing forces, where the Bill reflects a society attempting to navigate the space between innovation, human agency, and machine autonomy. Its significance lies not in offering final answers, but in acknowledging that governance in the age of artificial intelligence is an ongoing negotiation—one that must continually adapt to a future it cannot fully predict, let alone control.