The US government ordered Anthropic to bar foreign nationals from its two newest Claude models, forcing the company to switch them off for every customer worldwide. Washington has moved from counting AI layoffs to seizing the models themselves, and the lab it hit first is the one that asked to be regulated. Hollywood's directors settle, US states write the disclosure law Congress will not, and OpenAI heads for an IPO while losing money on every dollar.
Export law shut one AI lab while its rival stayed open; public capital is now asked to fund the gap.
At 5:21pm Eastern on Friday 12 June, Commerce Secretary Howard Lutnick ordered Anthropic to bar foreign nationals from Claude Fable 5 and Mythos 5. The only way to comply was to switch both off for everyone on earth.
United StatesOn 12 June 2026, Commerce Secretary Howard Lutnick ordered Anthropic to bar all foreign nationals from Claude Fable 5 and Mythos 5. Those models cannot check nationality mid-conversation, so Anthropic switched both off for every customer on earth.
Lutnick used the deemed-export doctrine, a law built for chips, to reach a deployed consumer AI product. OpenAI's rival model carried the same jailbreak and stayed on sale.
New York's A9581 passed both chambers on the strength of a Senate vote and now awaits Governor Hochul's signature. California cleared SB 951 through its labour committee 5-0 on Wednesday 10 June. Washington's own federal bill has not moved.
United StatesNew York's A9581 passed both chambers and awaits Governor Kathy Hochul's signature. The bill requires firms with 50 or more employees to file an annual report on how AI affects their hiring. California's SB 951 cleared a Senate committee 5-0 on 10 June.
Both bills answer a specific failure: New York's existing AI layoff-disclosure clause produced zero attributions from 162 companies covering 28,300 workers in its first year. States are filling the gap Congress has not.
The Directors Guild reached a four-year AMPTP deal on Tuesday 9 June. Directors won the right to treat AI footage like a camera shot, but secured no training ban and no residual, with a 24.4% health-plan rise as their largest concrete gain.
The Directors Guild of America reached a four-year deal with the studios' body on 9 June 2026. Directors won the right to treat AI-generated footage like camera footage, but secured no training ban and no royalty. Their main concrete gain was a 24.4% health-plan rise.
The guild closed in 29 days, fastest of three. Studios hold a settled template through 2030: AI permitted, workers keep nominal credit, no guild secured payment for AI-generated work.
EU Parliament committees backed the Digital Omnibus 93-4 on Tuesday 2 June, sending it toward adoption before 2 August. The binding duty on employers to ensure staff understand AI is downgraded to an unenforceable promise to support it.
EU Parliament committees approved the Digital Omnibus 93-4 on 2 June 2026, sending it toward adoption before 2 August. The package swaps the binding employer duty to ensure staff understand AI for an unenforceable promise to support that understanding.
For EU workers in AI-augmented jobs, the change removes the one statutory tool to demand an explanation from their employer. Binding AI employment rules under the AI Act arrive only in December 2027.
OpenAI confirmed its IPO filing on 9 June, targeting a September listing above $1 trillion. Reported figures show roughly $2bn a month in revenue against a projected $14bn operating loss for 2026, with no positive cash flow expected until about 2030.
United StatesOpenAI confirmed on 9 June 2026 that it filed a draft prospectus targeting a September listing above $1 trillion. Reported figures show roughly $2bn a month in revenue against a projected $14bn operating loss for 2026, with no positive cash flow until about 2030.
Goldman Sachs, Morgan Stanley, and JPMorgan took the underwriting mandate. OpenAI's March 2026 round valued it at $852bn on roughly $122bn from Amazon, Nvidia, and SoftBank, bridging losses to a projected 2030 breakeven.
Up. The Lutnick directive raises the ceiling for US government intervention in AI deployment from study-and-report to runtime withdrawal without notice. The next escalatory step is a second lab facing a BIS action, which would confirm the channel as a repeatable policy tool. The competitive-reporting mechanism, if it becomes known as a viable route to imposing costs on rivals, creates an incentive for further filings. On the capital side, OpenAI's S-1 disclosure will, for the first time, submit the displacement thesis to public-market due diligence: if analysts at Goldman Sachs, Morgan Stanley and JPMorgan conclude the 2030 breakeven is not credible, a valuation correction cascades to every private AI company that took funding at 2025-26 prices. The EU has no near-term escalation path on worker rights before December 2027.
