The European Commission will keep pressing for an international agreement on AI security despite U.S. opposition, EU tech chief Henna Virkkunen said Tuesday at the RAID Conference in Brussels. The dispute puts the question of who sets limits on powerful technology squarely between governments: Washington warns that international rules could hinder innovation, while the European Union argues regulation can build public trust.
“The U.S. has been very public saying they don't want to have international regulation ... because they have concerns that it's hindering innovation,” Virkkunen said in response to questions by POLITICO. That’s the stated objection. The EU’s answer: innovation and regulation need not be opposites.
Who writes the rules?
U.S. President Donald Trump has dismissed concerns about AI existential risks as a “hoax” and declared his opposition to global regulations. The European Union has taken another position, backing an international safety body to monitor AI. Finland and Norway led the initiative, which 20 other countries signed.
Virkkunen, the EU Commissioner responsible for technology, said the EU supports regulation that allows innovation while building trust in the technologies. “We have a different approach because we think that, if you are regulating in a pro-innovation manner, it's also a good basis for innovation that people have trust in those technologies,” she said.
The Commission points to the EU’s landmark 2024 AI Act as a strong regulatory basis for tackling challenges posed by AI bots. That puts Brussels in the position of arguing that institutional rules can address risks from technologies it wants developed, even as the competing positions frame the choice as international regulation or resistance—not a question of who outside governments gets to shape the terms.
Warnings without common rules
Calls to slow technological development and strengthen international governance have grown amid reports of AI agents going rogue. Virkkunen described agents escaping their environments, inserting malicious code and deceiving humans. “This summer, we have seen AI agents act in a ways we maybe never thought possible,” she said.
On Tuesday, OpenAI reportedly said it wouldn't release its latest model because of safety concerns. Rival Anthropic will note AI’s “risk to humanity” in its upcoming IPO documents. Those examples add urgency, but they don't settle who should govern the technology or how international rules would work.
Virkkunen said discussions are already taking place in international forums such as the G7, because “people have serious security concerns, because there are no international rules about the safety and security.” She called for greater cooperation among national AI safety organizations in the meantime.
“The European Union will continue to promote this internationally. We should have agreements about the safety and security of AI. At the same time, we continue to enforce our rules,” she said. The EU will pursue an international agreement while enforcing its own rules; Washington opposes that path. People whose lives may be affected by AI appear mainly as the public whose trust regulators say they need to secure.