Kenya Creates AI Approval System to Protect Citizens

Kenya is taking a bold step to ensure artificial intelligence serves people fairly, not just efficiently.

A new bill proposed by Senator Karen Nyamu would require companies to get government approval before using high-risk AI systems that make important decisions about people's lives. The law would cover AI tools that determine who gets loans, screen job candidates, diagnose medical conditions, and identify people through biometric data.

The proposal comes at the perfect time. AI adoption is surging across Kenya's tech sector, with startups using automated systems for everything from fraud detection to customer service. But when algorithms start deciding who deserves credit or employment, the stakes get much higher.

Under the draft Artificial Intelligence Bill 2026, companies would need approval from a new AI commissioner before launching these systems. The commissioner would maintain a public registry of AI tools in use and have power to inspect how they work. Companies that skip approval could face fines up to 5 million Kenyan shillings (about $38,000) or three-year jail terms for executives.

This makes Kenya a global leader in AI accountability. While the European Union and United Kingdom rely on audits and administrative penalties, Kenya is considering criminal consequences for harmful AI deployment. That sends a clear message: protecting people comes before moving fast and breaking things.

Mike Olukoye, a Nairobi tech legal expert, supports the oversight. He points out that AI systems already influence credit decisions, hiring outcomes, and access to essential services. Early intervention can prevent harm before it affects thousands of people.

The Bright Side

The law reflects a hopeful truth: technology doesn't have to be an uncontrolled force. By creating guardrails now, Kenya is showing other nations how to embrace innovation without sacrificing fairness.

The approval process might slow some product launches, but it could save countless people from algorithmic discrimination. A loan applicant denied by a poorly trained AI system loses more than time. They lose opportunity.

Kenya's approach also recognizes that AI isn't neutral. These systems learn from data that often reflects existing biases. Without oversight, they can automate and amplify unfairness at scale.

The public registry is particularly smart. Transparency helps people understand when AI is making decisions about their lives. It also creates accountability, encouraging companies to build systems they'd be proud to defend publicly.

Startups may worry about bottlenecks, especially those using third-party APIs from global providers. But the bill's focus on high-risk systems means routine automation can proceed without bureaucratic hurdles. The question becomes how broadly regulators will define "high-risk" in practice.

Companies will need to balance compliance with protecting their proprietary technology. The inspection powers could expose trade secrets, creating tension between transparency and competition.

Still, the bill represents progress worth celebrating. It shows a country choosing to shape its technological future rather than simply accepting whatever Silicon Valley exports.

Kenya is proving that developing nations don't have to be passive recipients of new technology, but can lead the conversation about how innovation should work for everyone.