European leaders at the G7 summit in France called on the US to share access to the latest artificial intelligence models amid growing concern over Washington's recent export controls.
The host, French President Emmanuel Macron, said the risk is that AI companies may lose clients if the US can suddenly "turn off the switch." Without transatlantic co-operation, "we will fracture the world, and we will have no effective solution", Mr Macron said.
A platform will be set up in the next month among a handful of Western democracies, with their leaders planning to meet in September, Mr Macron said. He had earlier convened a working lunch in the presence of G7 leaders and big-tech chief executives.
At the lunch, European Commission President Ursula von der Leyen said it was in the “mutual interest” of US and European companies to share technology.
“We use each other's trusted technology, and our financial systems are interconnected. It is in our mutual interest that our citizens and companies can safely use the best AI models,” Ms von der Leyen said.
“We have complementary strengths,” she added. “Frontier models developed by the US have unprecedented capabilities, and the EU is developing a vibrant innovators ecosystem and becoming a leading market for the adoption of industrial AI.”
AI company Anthropic's chief executive Dario Amodei, who held a bilateral meeting with G7 host French President Emmanuel Macron, was the focus of attention after his firm on Friday disabled access for all users to Claude Fable 5 and Mythos 5.
This came after US President Donald Trump ordered Anthropic to block foreign nationals from accessing its most advanced AI models, citing national security concerns. In a social media post, Mr Trump called the firm a “radical left” and “woke” company.
But Europeans were expected to avoid discussing the feud and focus instead on boosting cross-Atlantic collaboration on AI, by lobbying for Europe to be considered a trusted partner by the US.
Mr Trump sat at a table between Open AI's Sam Altman and Google's DeepMind chief executive Demis Hassabis at the working lunch. Diplomatic sources described conversations over the sharing of the most advanced AI models as “delicate and evolving”.
Trusted partners
The EU as a bloc is also launching initiatives to strengthen its own AI investments, but continues to lag behind the US. German Chancellor Friedrich Merz said the potentials of these new technologies should be available for all countries, but that Europe also needed to catch up.
“I very much hope that we will have an intensive co-ordination with the American government, because the potential of these new technologies should be available for all countries,” Mr Merz said on Wednesday.
The so-called “trusted partners” could be countries or companies, Reuters reported. An agreement providing broader access to advanced models would allow G7 countries to use the models to develop stronger cybersecurity defences against rivals such as China.
“If you don't want China to catch up on US and European tech, then let's set up trusted partners to exchange innovations,” the sources said.
In total 12 big-tech leaders were invited to the G7, including Meta's Alexandr Wang. Several of them held bilateral meetings with Mr Macron. All the countries of the G7 are to be represented, with chief executives from France – Mistral AI's Arthur Mensch – and the UK – Synthesia's Victor Riparbelli. India, a G7 guest this year, is to be represented by Vivek Raghavan, founder of Sarvam AI.
Speaking to Politico, European tech leader Domyn's Uljan Sharka said he did not blame Mr Trump for his move targeting Anthropic AI. “They were pushed and forced to take action,” he said, referring to Anthropic’s branding of its models as highly capable of finding software vulnerabilities.
Anthropic's Mythos, developed to find flaws in computer code to help bolster defences against cyberattacks, is seen by cybersecurity experts as capable of turbocharging attacks on the technology systems of banks, which it aims to protect.
Child safety
Dangers posed by AI-generated deepfakes and social media to children and teenagers across the globe are also on the G7 agenda. In a statement, leaders said “digital service providers have an important role and opportunity to provide digital platforms that are safe by design, secure, privacy-preserving, age-appropriate and protecting of children and youth, including by default settings”.
Making conversational AI safer is among the top list of priorities for G7 leaders as states toughen their positions on child access. “Recommendations systems should, when used, be designed to elevate age-appropriate content and to reduce exposure to risks,” the statement said.
UK Prime Minister Keir Starmer earlier this week announced a social media ban for under-16 year-olds. This includes popular apps such as Instagram, TikTok and Snapchat. In France, Mr Macron has vowed to introduce a similar ban for children under 15 before the end of the year.
