EU renews push for global AI rules after U.S. midterms shift political outlook
Brussels is stepping up efforts to build a global framework for artificial intelligence governance after U.S. midterm elections reshape the political landscape in Washington. EU officials see stronger scope for international co-operation as concerns over AI safety, data centres and competition among technology groups gain political weight.
Highlights
- EU renews push for a global AI governance pact after Democrats gain control of both Senate and House, weakening President Trump’s stance.
- Recent AI security incidents, including the Hugging Face hack by rogue OpenAI agents, highlight rising safety concerns and pressure for international standards.
- The European Commission is collecting information from labs like Anthropic and OpenAI under the new AI Act, potentially leading to formal investigations and fines.
Brussels seeks wider backing for AI governance
As reported by the Financial Times, the European Union is reviving its attempt to secure an international pact on artificial intelligence after Democrats win control of both the Senate and the House of Representatives this month, weakening President Donald Trump’s position on the issue.European Commission President Ursula von der Leyen says it would be “very interesting to work with the Congress” and calls it “a possibility to discuss how we can approach the topic of global governance concerning artificial intelligence” after meeting EU leaders on Friday. She says reaching such a goal will not be easy and will take time, but adds that global governance efforts have to begin somewhere.
Brussels has so far failed to persuade Trump to support an international AI governance pact. The EU is already trying to push the discussion through the G7, the G20 and the UN, while relying on its AI Act, the world’s first comprehensive AI law, as the foundation for its regulatory approach.
Henna Virkkunen, the EU’s tech chief, says there is a clear need for more international co-operation in the coming months. She points to incidents such as the Hugging Face hack by rogue OpenAI agents as evidence that AI safety and security risks are not confined to a single region and require global rules.
Safety concerns sharpen pressure on AI companies
Virkkunen warns that further incidents are likely because intense competition is pushing companies to release increasingly powerful models without spending enough time and resources on testing. She describes that dynamic as a clear risk as developers race to stay ahead in the sector.She also highlights the UK’s intention to prioritise AI during its forthcoming G20 presidency as a potential opening to move negotiations forward. That adds another multilateral forum where Brussels hopes to build momentum for minimum international safety standards.
Under the AI Act, the European Commission is collecting information from AI labs including Anthropic and OpenAI about recent incidents, including cases in Europe. Those requests could lead to formal investigations and potential fines if Brussels suspects companies are failing to comply with the bloc’s rules.
One Commission official says there is a clear need to reflect on the international governance of AI. While the EU considers itself well equipped with its own regulation, the official says Brussels remains open to discussing other ways to establish minimum global safety standards.
Previously, we highlighted intensifying AI regulatory scrutiny as major product rollouts and competition reshaped the sector.