The head of the Council of Europe, Alain Berset, made his inaugural visit to Washington last week, aiming to raise the profile of an international AI treaty that the Biden administration had endorsed but that remains largely unknown outside European circles. Berset, a former Swiss president who now leads the Strasbourg-based organization of 46 member states, met with lawmakers, think tanks, and journalists to promote the Framework Convention on Artificial Intelligence.
The treaty, finalized in September 2024, was signed by the Biden administration along with more than 20 other countries, including Japan, Israel, and several European and American nations. It is a legally binding document, but so far only the European Union has ratified it, and that happened as recently as May 2026.
Berset's push comes at a time when President Trump has rejected and criticized AI regulation, which he says hampers innovation. This stance presents a major obstacle to international cooperation on AI safety, a concern that has been magnified by warnings from tech leaders about the technology's potential to cause catastrophic harm.
A 'hard wake-up call'
In an interview, Berset stressed the urgency of state involvement in AI governance. "I do not see any alternative to a strong, I would say, involvement of states," he said, when asked about the challenge of rallying cooperation amid a global AI arms race. He argued that the framework, already drafted and agreed upon, provides a concrete starting point.
"For the first time, we have specialists telling us that this revolution could destroy humanity in a few years," Berset said. "This should be a quite hard wake up call for all of us, to say, 'OK, what do we propose for this?' And the framework convention is exactly that."
His Washington visit followed a stop in New York, where the Council of Europe opened a permanent observer office at the United Nations. The office aims to boost the council's visibility and promote its work on human rights, democracy, and the rule of law, with AI as a dominant topic. The U.N. provides a global platform for the council to advance its AI framework, which is grounded in those principles.
"Our task is to connect these strengths," Berset said at the opening of the council's U.N. office.
A legacy of treaties
The Council of Europe, founded in 1949 in the aftermath of World War II, is one of the continent's oldest human rights bodies. It produced the European Convention on Human Rights, enforced by the European Court of Human Rights, and has drafted roughly 200 treaties. Berset noted that the court has issued one million decisions, making the council a valuable resource for justice and accountability.
Unlike the European Union, the council's membership extends beyond EU countries. Its 46 members include aspiring EU members like Turkey and Ukraine, as well as observer states such as the U.S., Canada, Mexico, Japan, and the Holy See. This broad membership, Berset said, creates "the biggest common legal space that exists in the world," though it is not always visible.
"It's not that visible but it's extremely important for having this convergence, year after year, between the 46 members of the Council of Europe," he said in a speech at the Hudson Institute.
Berset acknowledges that the framework is just one contribution to AI governance. "I think it is one contribution. It is not the only one, it will not be the only one, but it is one concrete contribution," he said. To critics, he challenges them to offer better solutions: "I would be more than happy if you bring a better answer. We have at least one answer, we have one element. We have one framework where we are able to work together to follow this."
The push for international AI rules comes as a UN-backed panel warns that AI safety measures are failing, and while partisan warfare over AI regulation leaves safety in limbo in Washington. Berset's visit underscores the growing divide between those seeking binding international standards and the current U.S. administration's preference for deregulation.
