EU Outpaces US on AI
The EU is moving decisively: its AI Act will come into force in August 2026 and has extraterritorial reach—any company whose AI is used by EU residents must comply, regardless of where it’s based, forcing foreign vendors to adjust product and legal plans (nadcab.com). U.S. lawmakers are publicly struggling to keep pace, and Kenya is simultaneously advancing its own comprehensive Artificial Intelligence Bill 2026—creating a patchwork of high‑stakes rules that will raise compliance costs and strategic choices for tech firms (washingtonpost.com) (hapakenya.com).
The AI Act formally became EU law on August 1, 2024 (commission.europa.eu), but its most sweeping application date—when the bulk of high‑risk and provider obligations apply—is scheduled for August 2, 2026 (programming-helper.com). The regulation builds in tiered penalties, including fines up to €35 million or 7% of annual global turnover for the gravest breaches and up to €15 million or 3% for other violations, with smaller tiers for misleading information; those amounts are specified in the Act’s penalties article. (artificialintelligenceact.eu) The law explicitly covers any AI “placed on the EU market or used” in the EU—meaning providers anywhere must meet EU rules if their systems interact with EU residents—and the EU already phased in several obligations for general‑purpose models and bans earlier this year. (unorma.com) U.S. lawmakers have yet to pass a comprehensive parallel statute, a gap highlighted in a Washington Post piece that includes an interview with Rep. Sam Liccardo on congressional shortcomings, while administration actions have favored export‑control and agency rules over a single congressional framework. (washingtonpost.com) Consultancy and policy analyses warn of heavy compliance bills for smaller firms: one estimate suggests initial compliance costs for a 50‑person AI company could run between €320,000 and €600,000 with recurring annual costs near €150,000, and broader estimates put sectoral impacts in the billions. (theparliamentmagazine.eu) Kenya’s Artificial Intelligence Bill, 2026—sponsored by Nominated Senator Karen Nyamu—would create a national AI regulator (the Office of the Kenya AI Commissioner) and proposes penalties such as fines of up to KSh5 million or potential jail terms for misuse, signaling another major jurisdiction imposing bespoke rules. (new.kenyalaw.org) Global vendors are already weighing tactical responses: major consultancies and auditors are advising firms to build EU‑specific compliance programs or separate product builds to meet risk‑tiered documentation and conformity‑assessment requirements in the short term. (kpmg.com)
