AI Governance 2026: The Race Between Rules and Capabilities
By Sanna the Weaver • Sat Feb 21 2026 • Technology
On August 1, 2024, the European Union's AI Act entered into force — the world's first comprehensive legal framework for artificial intelligence. By March 2026, with the regulation's most significant provisions for high-risk AI systems now fully applicable, its effects are becoming visible: some AI products have been modified or withdrawn from the EU market, compliance costs have risen for AI developers operating in Europe, and regulators are beginning to investigate their first cases. Meanwhile, the United States has no equivalent federal framework, and China operates under a patchwork of sector-specific rules. The result is a fragmented global AI governance landscape that reflects the competitive dynamics of the technology it is meant to govern. The EU Act's First Real Test The AI Act categorizes AI systems by risk level — unacceptable risk (banned), high risk (strictly regulated), limited risk (transparency requirements), and minimal risk (no specific obligations). Facial recognition in public spaces, social scoring systems, and AI that exploits psychological vulnerabilities fall in the banned category. High-risk applications — including AI in hiring, credit scoring, medical devices, and critical infrastructure — must undergo conformity assessments, maintain detailed documentation, and provide human oversight mechanisms. The first major enforcement cases are expected in 2026 as regulators build their investigative capacity. America's Non-Framework The United States has no federal AI law. The Biden-era executive order on AI safety, which required developers of frontier models to share safety test results with the federal government, was revoked by the Trump administration in January 2025. The administration's position is that regulation would harm US competitiveness and that industry self-governance is sufficient. Critics note that industry self-governance produced the social media environment that Congress is still trying to regulate a decade later. "When scientific AI capability accelerates, governance must learn to move earlier — not just faster." — GESDA Science Breakthrough Radar, 2026 The US-China AI Race and Its Governance Dimension US-China competition for international AI markets will heat up significantly in 2026. The governance dimension of this competition — which country's AI standards, liability frameworks, and deployment norms become the global default — is as consequential as the technological dimension. Countries that adopt Chinese AI systems adopt Chinese data practices and Chinese surveillance capabilities alongside them. Countries that adopt American AI systems face a governance vacuum, since the US has not articulated the rules it expects its AI to operate under. The 2026 GESDA Science Breakthrough Radar, distilling insights from nearly 2,400 leading researchers, concludes that the governance gap is the single greatest risk in AI's trajectory — not the capabilities themselves, but the absence of frameworks to guide their use.