September 15, 2026, 17:44 GMT | Comment
AI companies' calls to slow frontier-model development could help the EU extend the AI Act's global influence, echoing the "Brussels effect" created by the GDPR. But the
European Commission has been reluctant to engage on the issue and has largely stressed enforcement in its own backyard. For some policy experts, this is a missed opportunity.
Calls from artificial intelligence companies to slow down the development of frontier AI models may give Brussels a fresh opening to revive the "Brussels effect" for the EU AI Act, but the European Commission doesn't seem to be actively pursuing it.
Over the weekend,
Anthropic CEO Dario Amodei proposed "pacing the frontier" of AI — which will require the involvement of governments and global cooperation — citing risks of loss of control over AI systems, cyberattacks, bioterrorism and serious economic disruption.
The current debate over frontier risk is the perfect opportunity to trigger the Brussels effect for the AI Act, according to Kai Zenner, head of the office of
European Parliament lawmaker Axel Voss. The term describes the EU's ability to project regulation beyond its borders, as it has with its data regulation law, the GDPR.
Jimmy Farrell, a policy expert at Brussels-based think tank Pour Demain added that Europe could use this moment "to promote the fact that the EU has moved on this first."
"The EU is already furthest in the world in terms of a regulatory framework to address frontier AI risks," Farrell said.
Almost every AI provider has signed the General-purpose AI, or GPAI, Code of Practice, which lays out safety obligations for AI model providers and could lay the groundwork for some international governance regime.
There are already some signs that the world is following the EU's lead on AI. The US has seen state legislation on frontier AI in California, New York and Illinois that is close to the AI Act's GPAI provisions.
But unlike the GDPR, which generated a Brussels effect as foreign companies adapted to Europe's stringent data-protection rules, the AI Act may face a more difficult path to similar influence. China and the US are building their own approaches to AI, and national security considerations heavily shape the debate. For them, falling behind in AI means risking both cybersecurity and economic disadvantage.
Europe's response to Amodei's warnings over the catastrophic dangers of AI, including calls for international governance, have also been tepid compared to reactions from lawmakers in Australia (see
here), the US (see
here) and other jurisdictions.
Responding on Monday, a commission spokesperson reiterated the same response it has given over the past several weeks: the AI Act is being enforced and developers will have to prove their services are safe to use. More than 30 companies have already been served with information requests (see
here).
Although EU tech sovereignty commissioner Henna Virkkunen previously called for international cooperation on AI risks, no concrete moves have been announced (see
here).
Brussels seems to be giving up an opportunity to lead "a third way" on AI even though international partners are interested in collaborating, Zenner said. Commission services and their counterparts could work on things such as regulatory sandboxes, model evaluation or incident reporting.
Other policy observers say AI companies' warnings of catastrophic risks do not justify urgent government action.
The EU should resist "scare stories coming out of Silicon Valley, however alluring, and focus on actual adoption risks and rewards," said Patrick Grady, head of European Tech at the Information Technology and Innovation Foundation, a public policy think tank.
The warnings from tech executives could as well be an attempt at regulatory capture, in which Anthropic and
OpenAI could use fears of AI to influence the terms and teams that regulate their industry, he added.
For Brando Benifei, the European lawmaker who led negotiations on the AI Act, the immediate priority is enforcement of Europe's existing framework rather than tailoring regulation to the latest interventions by Silicon Valley executives.
"We cannot calibrate regulation or enforcement to whatever Amodei, Altman or Musk says on a given day," Benifei said. "Europe has created a legal framework, with a public watchdog and enforceable obligations, and the priority now is to use it."
The international cooperation on frontier AI could look more like the slow-burn process to prevent the proliferation of nuclear weapons. These are already ongoing with initiatives such as the Hiroshima AI Process, which hopes to establish international rule-making for advanced AI systems, or the OECD's Global Partnership on Artificial Intelligence, Benifei noted.
Whether this would be fast enough to prevent the catastrophic risks of frontier AI, however, is far from clear.
Please email editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers.
Tags
Sections:
Artificial Intelligence
Industries:
Computing & Information Technology
Geographies:
Europe, European Union Member States