AI Governance Shifts as Nations Address Safety and Oversight

International efforts to manage artificial intelligence have intensified, with multiple governments and organizations introducing new safety reports, regulations, and policy frameworks since 2023.

The First Independent International AI Safety Report, led by Yoshua Bengio and commissioned by 30 countries, was released on 29 January 2025 following the 2023 AI Safety Summit at Bletchley Park. This report was followed by a second edition published on 3 February 2026 ahead of the AI Impact Summit, reflecting continued international engagement with AI safety issues.

The International AI Safety Report highlighted risks such as unpredictable AI behavior, unreliable reasoning, and objectives that may not align with human oversight. The report also identified systemic concerns, including reliance on a limited number of model providers, risks of interconnected infrastructure failures, and challenges to institutional stability.

Authors of the report stated that AI developers often lack comprehensive understanding of their models, which complicates the management of associated risks. The report further noted that advanced AI systems could undermine human oversight either by design or through lapses in attention.

What the numbers show

• AI-related legislative mentions rose by 21.3% in 2025 across 75 countries compared to 2023
• U.S. federal agencies introduced 59 AI-related regulations in 2024, more than double the previous year
• The first International AI Safety Report was published on 29 January 2025

Between 2023 and 2025, oversight of AI was primarily managed by industry through voluntary measures. By early 2026, government agencies began to play a larger role, including the U.S. Food and Drug Administration issuing guidance for AI-enabled medical devices and the National Institute of Standards and Technology developing risk assessment standards.

On 23 July 2025, the White House published “Winning the Race: America’s AI Action Plan,” outlining over 90 federal actions intended to strengthen U.S. leadership in artificial intelligence. This policy blueprint was part of a broader trend toward increased government involvement in AI governance.

In January 2025, the U.S. Bureau of Industry and Security released the Framework for Artificial Intelligence Diffusion, an interim final rule that was scheduled to take effect in May 2025. However, the rule was rescinded before implementation on 13 May 2025.

Industry reaction

The U.S. AI diffusion export control rule was criticized by some institutions for potentially undermining national leadership in AI, according to statements documented at the time. The decision to rescind the rule before it took effect was confirmed by the Bureau of Industry and Security.

International organizations and governments have continued to publish reports and policy documents to address the evolving landscape of AI safety and governance, reflecting a shift from voluntary to regulatory oversight in recent years.

* This article is based on publicly available information at the time of writing.