Story perspectives

G7 Weighs US AI Access; Anthropic Ban Raises Banking Risks

6/17/2026

Summary

Story summary

G7 leaders floated a “trusted partners” plan to let non-U.S. entities use U.S. frontier AI models despite the ban.
U.S. Commerce Secretary Howard Lutnick met G7 delegates at the Evian-les-Bains summit to discuss it, with no announcement expected Wednesday.
Washington halted access to Anthropic’s models, prompting cybersecurity experts to warn that its Mythos model—sought by the EU for assessment—could accelerate banking attacks, though European banks have not deployed it.

History

Historical Context

During the Cold War the U.S. and allies created CoCom in 1949 to license critical technology, placing encryption on the Munitions List in 1954. Legal challenges and the 1996 Executive Order moved commercial cryptography to the Commerce Control List, ending the 1990s “Crypto Wars.” Today G7 leaders cite that precedent while debating AI export limits.

1 of 2

Story summary

G7 leaders floated a “trusted partners” plan to let non-U.S. entities use U.S. frontier AI models despite the ban.
U.S. Commerce Secretary Howard Lutnick met G7 delegates at the Evian-les-Bains summit to discuss it, with no announcement expected Wednesday.
Washington halted access to Anthropic’s models, prompting cybersecurity experts to warn that its Mythos model—sought by the EU for assessment—could accelerate banking attacks, though European banks have not deployed it.

1 / 2