Macron pushes for US access to Anthropic AI tools over safety worries

His comments follow a pledge by G7 leaders for closer coordination on both the opportunities and risks presented by frontier AI, News.Az reports, citing The Independent.

The leaders, who convened for the Group of Seven summit in Evian-les-Bains, France, discussed the establishment of a "trusted partners" scheme. This initiative aims to grant non-US nations access to advanced AI models from firms such as Anthropic.

However, cybersecurity experts have voiced concerns over Anthropic's Mythos. While developed to identify coding flaws and bolster cyberdefences, there is a potential risk that the tool could inadvertently be used to accelerate attacks on the very systems it is designed to protect.

Last week, U.S. President Donald Trump told Anthropic to block foreign nationals from accessing its advanced models, citing national security concerns. That move spurred G7 discussions on the creation of the "trusted partners" scheme, which could open a path around the U.S. restrictions.

The "trusted partners" could be countries or companies, Reuters reported on Tuesday, and would allow them to use the models to develop stronger cybersecurity defences against rivals like China.