Europe moved the dates of its AI rules. Why a developer in Kenya should care
The EU AI Act is the world's most detailed law on artificial intelligence. In 2026 its timetable changed. Here is what moved, what did not, and why it reaches beyond Europe.
By Howell Asava, Developer, Vukalink ยท
Laws written in Brussels have a habit of becoming the rules everyone else follows. Data protection went that way: Europe's GDPR shaped privacy laws in many countries, Kenya's among the influences. Artificial intelligence may go the same way, which is why I follow the EU AI Act even though I build software in Kenya.
The law in one paragraph
The AI Act entered into force on 1 August 2024 and applies in stages. It sorts AI systems by risk. A few uses are banned outright. "High-risk" uses, such as AI that helps decide who gets a job, a loan, a place at school or access to public services, must meet strict requirements for risk management, documentation and human oversight. Most everyday AI faces lighter duties, chiefly to be open about the fact that it is AI.
What moved in 2026
The technical standards needed to show compliance for high-risk systems were not ready in time. So the EU adopted a simplification package, known as the Digital Omnibus on AI, which the Council approved in late June 2026. Its main effect:
- Obligations for stand-alone high-risk systems moved from 2 August 2026 to 2 December 2027.
- High-risk AI built into regulated products, such as medical devices and machinery, moved to 2 August 2028.
What did not move
The transparency duties in Article 50 still apply from 2 August 2026. In plain terms: people must be told when they are dealing with an AI system, and AI-generated content must be marked as such. Systems already on the market before that date have until 2 December 2026 for the technical marking of generated content.
Why it matters outside Europe
The Act applies to anyone who offers AI systems to people in the EU, wherever the company is based. A Kenyan startup with European customers is in scope. And even for those without them, the Act is becoming a reference point: investors, large customers and other governments borrow its vocabulary.
My own practice, which I think costs little and builds trust: say clearly when something is AI and when it is not. The question box on this website, for example, is labelled as answering from the site's own information and not from a live AI model, because that is what it does.
This is an explanation, not legal advice. If your product may fall under the Act, ask a lawyer who practises in this area.
References
- European Commission AI Act Service Desk, Timeline for the implementation of the EU AI Act
- Gibson Dunn, EU AI Act Omnibus Agreement: Postponed High-Risk Deadlines and Other Key Changes (27 May 2026)
- Cloud Security Alliance, EU AI Act High-Risk Deadline Pushed to December 2027 (8 July 2026)
- artificialintelligenceact.eu, High-level summary of the AI Act (updated August 2026)
Numbers quoted above are as reported by these sources; the opinions are the author's own.