Canada's Carney warns U.S. AI limits on Anthropic show risk; UC Merced uses it
Canadian Prime Minister Mark Carney said Sunday that U.S. curbs on Anthropic’s newest AI models show the danger of relying on a few providers. UC Merced lists Anthropic among supported tools, a local reminder to plan for supply shocks.
Canada's Carney warns U.S. AI limits on Anthropic show risk; UC Merced uses it
Key Takeaways
- On June 14, Canadian Prime Minister Mark Carney warned U.S. limits on Anthropic’s latest AI models show the risk of reliance on few providers.
- Anthropic took its new Fable 5 and Mythos 5 models offline for foreign users to comply with a U.S. directive.
- UC Merced’s IT guidance lists Anthropic among supported tools, so policy swings can ripple into Valley classrooms and labs.
- AI policy will be discussed at the G7 this week, with no Carney–Trump bilateral meeting scheduled.
Speaking in Westport, Ireland, on Sunday, Canadian Prime Minister Mark Carney said U.S. restrictions on Anthropic’s newest AI models reveal the risk of leaning on a small number of American providers. That matters here. UC Merced lists Anthropic among supported tools in campus AI resources, and Fresno State is building out an AI campus effort of its own.
Which is the point.
What Carney said
Carney pointed to Anthropic’s move to take its new models, Fable 5 and Mythos 5, offline for foreign users after a U.S. directive, and he urged countries to diversify their AI options. He tied the model pause to a broader lesson about supply risk, comparing it to depending on a single vendor for key infrastructure. "It is never a good idea to have one option," he said.
Why the Valley should care
Local access to Anthropic for domestic use isn’t what Washington targeted, but cross‑border projects can get sticky overnight. UC Merced’s guidance notes Anthropic among tools available inside Zoom’s AI features and through Amazon’s AI services, so a campus lab working with a Canadian partner could suddenly need a workaround if access terms shift mid‑semester. Fresno State’s AI initiative points students and faculty to centralized, supported tools as well. Contracts help, until policy changes faster than the paperwork.
Outside, Fresno hit triple digits by midafternoon.
Valley firms that buy AI‑enabled software, from office copilots to crop‑planning dashboards, often depend on the same handful of U.S. model providers behind the scenes. If a vendor quietly swaps models or disables features to stay inside federal rules, the change can surface here as slower workflows or missing buttons, not a headline. The practical question for local IT leads and department chairs is simple and unglamorous: do they have a second provider ready, and do their data rules cover that pivot.
The policy backdrop
Carney spoke on June 14 ahead of AI talks at the G7 in France. He said trade and tech diversification remain priorities for Canada while free‑trade review discussions proceed without a leader‑to‑leader meeting on the calendar. For Central Valley institutions, the moving parts to watch are U.S. export and access rules for frontier models, plus any vendor notices about feature changes for cross‑border users. What hasn’t been answered yet is whether Washington will narrow or broaden these limits as talks unfold.
"You’ll hear me say this over and over again. It is never a good idea to have one option."
Central Valley AI is produced by the CVAI Newsdesk team and developed by Kaweah Tech, a regional firm that builds, deploys, and integrates AI solutions for businesses across California's Central Valley.