Only 6% of Americans Use Claude, Anthropic Is at War With the Government, and AI Leaders Hit G7

Today we're covering the biggest AI stories of June 18th, 2026. New details on the Anthropic-U.S. government standoff are spilling into the open — a leaked letter from Commerce Secretary Howard Lutnick warned Anthropic against distributing Mythos to foreign persons, internal employee messages obtained by the New York Times show staff calling the situation unfair targeting based on "bad vibes," and the Washington Post reports the list of companies with Mythos access had quietly ballooned to include a South Korean firm with suspected ties to China — all while Dario Amodei and other AI CEOs are at the G7 summit in France proposing a U.S.-led international AI coalition. Pew Research published its 2026 AI survey of 5,000 U.S. adults showing chatbot usage just crossed 50% for the first time and a quarter of Americans now use one daily — but nearly 40% expect AI to make society worse over the next 20 years, with trust continuing to slide even as adoption climbs, and Claude registering at just 6% awareness against ChatGPT's 44%. Plus, Anthropic released a study analyzing 400,000 Claude Code sessions finding that your domain expertise matters more than your coding skill — lawyers, scientists, and managers with zero coding background nearly matched software engineers — and today's community workflow comes from a retired birder with no software background who built a full web app called Chase Report that reads eBird checklists, generates field summaries, and maps rare bird sightings across multiple states — all by describing what he wanted in plain English.