Anthropic CEO Dario Amodei told G7 leaders, including US President Donald Trump, that they must "resist the temptation to splinter" regarding the deployment of advanced AI tools. This call for unity came amidst international concerns about access to cutting-edge American AI models.
World leaders like French President Emmanuel Macron and Indian Prime Minister Narendra Modi voiced worries that the US could abruptly cut off their countries' access to top American AI models at any time. Macron specifically warned that if the US could "from one day to the next turn off the switch," it would adversely impact European economies and damage AI firms themselves. This sentiment was echoed by Prime Minister Modi, who emphasized the importance of unfettered access to leading AI models for democratic nations to safeguard critical infrastructure.
The concerns followed a recent decision by the US government to block Anthropic from exporting its newest Mythos 5 and Fable 5 models, citing national security grounds, after Amazon reportedly alerted the White House about potential bypasses of safety guardrails. Cybersecurity experts noted that similar capabilities exist in other freely available models, including those from OpenAI. This incident highlighted the risk for international companies and governments relying on US AI infrastructure, as access could be revoked without explanation.
In response to these anxieties, G7 leaders discussed creating a "trusted partners" scheme to grant non-US nations access to advanced AI models from companies like Anthropic and OpenAI. The goal is to establish an open trade network that bypasses US restrictions, allowing both countries and companies to become trusted partners if they use the models to bolster defenses against rivals like China. Macron suggested that it would be in Washington's interest to support such a scheme to ensure broader access to Mythos, as potential clients would be deterred from purchasing US AI if there was a risk of it being shut off suddenly.
During a working lunch at the G7, attended by AI executives including Anthropic's Dario Amodei and OpenAI's Sam Altman, the discussion revolved around regulation and AI infrastructure. OpenAI CEO Sam Altman urged the G7 to take charge of AI governance, stating that while AI labs develop the technology, democratic governments should be the ones to set the rules. European Commission President Ursula von der Leyen highlighted the mutual interest for the EU and US in the EU utilizing the best AI models, commending US efforts in ensuring AI firms act responsibly.
