The EU AI Act and Its Influence on AI Governance

© European Union, 2026, CC BY 4.0, via Wikimedia Commons

By MJ Weaver

In the global race for artificial intelligence (AI) dominance, the United States and China are competing for the leading technologies. AI advancement raises serious concerns about protecting humanity and the environment, yet it also holds the potential to improve healthcare, education, scientific research, and much more. As a result, there is a growing need for AI governance to address the risks of development. The European Union (EU) has come to the forefront of this effort with its AI Act, the first comprehensive regulation on AI worldwide. The AI Act entered into force in August 2024, and enforcement is shared among several national and EU-level bodies. The original compliance deadline was August 2, 2026, and non-compliant entities face heavy fines. However, compliance deadlines have been extended for high-risk AI until December 2027, and new regulations continue to be added.

Despite trailing the United States and China in AI capability, the EU is setting the global standard for AI regulation through the AI Act. Lawmakers signed the AI Act in June 2024, creating a framework for classifying AI systems through a risk-based approach. Systems with “unacceptable risk” are prohibited. “High-risk” systems are allowed but must meet a set of obligations and standards to access the EU market. Transparency-risk systems must meet disclosure requirements, while minimal-risk systems face no further obligations. A system is considered high-risk if it meets one of two conditions: it is a safety component of a regulated product covered by existing EU safety legislation, or it has a high-stakes use case. These use cases include biometric identification and categorization, critical infrastructure management, education and vocational training, employment, access to essential services, law enforcement, migration, asylum, and border control, and justice and democratic processes. Transparency-risk systems are those with the potential to impersonate or deceive, and they are subject to information and transparency requirements. Unacceptable-risk systems violate EU fundamental rights through manipulative, exploitative, discriminatory, or intrusive practices.

In addition to its risk-based rules, the act sets specific requirements for general-purpose AI (GPAI) models. GPAI models with “high-impact capabilities,” meaning capabilities that could pose systemic risk and affect the internal market, are subject to rigorous requirements. GPAI providers must prepare technical documentation for authorities, implement a copyright-compliance policy, and publish a summary of training data content. The act also requires every EU member state to establish at least one regulatory sandbox at the national level, which facilitates the development and testing of innovative AI systems under strict regulatory oversight.

Although the EU AI Act was the first piece of binding AI governance, the need for oversight is growing, and other countries are responding. In the United States, there is no federal AI law, but California, Colorado, Texas, and Illinois have enacted state-level laws on AI governance. Across Asia, countries are adopting frameworks, including South Korea’s AI Basic Act, China’s generative AI security standards, and Singapore’s leading initiatives on agentic AI, while Japan and India have released AI governance guidelines. Elsewhere, the United Kingdom’s Frontier AI Bill gave its AI Security Institute statutory powers, and Australia developed a national AI plan and an AI Safety Institute.

The EU AI Act bars noncompliant companies from entering the EU market, which pushes private entities to conform without any legislation being imposed on a global scale. Companies are beginning to cite the Act to inform their internal policies, illustrating its influence and private applicability. Multilateral institutions have also begun to adopt similar frameworks, such as the UN Global Dialogue on AI Governance, the International AI Safety Report 2026, and UNESCO’s Global Forum on the Ethics of AI.

Taken together, these developments show that the EU AI Act is shaping AI governance well beyond Europe’s borders. Because companies must comply to reach the EU market, the Act sets out a practical standard that many organizations adopt regardless of where they operate. At the same time, the extended deadlines and the ongoing additions to the regulation show that AI governance remains a work in progress. As the United States, China, and other nations continue to develop their own approaches, the challenge will be balancing innovation and competitiveness with protections for people, privacy, and the environment.