Skip to content
Daily Brief News Daily Brief

4 AI stories from June 13, 2026: US shuts down Anthropic's top models over a jailbreak claim, Google sues Chinese AI fraud ring, Meta's data-labeling revolt, and SpaceX closes IPO up 19%

· by Pondero Newsdesk · 4 stories

AI news daily brief: 2026-06-13

Four stories today: a regulatory first that shut down two Anthropic models worldwide, a major Google civil suit against an AI-powered fraud network, an internal workforce crisis at Meta, and SpaceX's record-breaking IPO debut.

US government shuts down Anthropic's Fable 5 and Mythos 5 worldwide, citing a narrow jailbreak

The US government delivered an export control directive to Anthropic at 5:21 PM ET on June 12, ordering the immediate global shutdown of Claude Fable 5 and Claude Mythos 5. Fable 5 had been publicly available for roughly three days. Anthropic complied within hours, then published a detailed rebuttal disputing the technical basis of the action.

The government's stated concern was a claimed jailbreak of Fable 5. Anthropic said it received no disclosure of a "universal jailbreak" and that the technique described, prompting the model to read a codebase and identify software flaws, is already available in OpenAI's GPT-5.5 and other publicly deployed models. The company argued its defense-in-depth classifiers remain active even when a narrow bypass exists in the model itself, and it characterized the shutdown as a misunderstanding of how its safety architecture works.

Anthropic said it is working to restore access and will share additional technical detail. The immediate question is whether the company obtains an emergency court stay in the next 24 to 48 hours. This is the first reported case of a US export control directive targeting a commercial AI model for immediate global shutdown. Full coverage: US government shuts down Anthropic's Fable 5 and Mythos 5

Google sues Chinese AI-powered cybercrime ring 'Outsider Enterprise' after 2.5 million scam texts in two weeks

Google filed a civil lawsuit on June 12 against a Chinese cybercrime network it calls "Outsider Enterprise," alleging the group used AI to send 2.5 million scam text messages impersonating Google and other brands to Android users over a two-week period. The operation deployed 9,000 fake websites and roughly one million fraudulent web domains.

Google said AI-powered detection tools intercepted more than 10 billion scam messages per month and that the FBI, AT&T, T-Mobile, and Verizon cooperated to block the network. Outsider Enterprise's phishing software, priced at $88 per week or $200 per month, has enabled an estimated $1.9 billion in losses and the theft of roughly 3.87 million credit cards since July 2023, per Google's complaint.

The lawsuit pairs criminal referral with civil action and carries the cooperation of three major US carriers, a pattern the company will likely use as a model for future AI-enabled fraud enforcement. Full coverage: Google sues Outsider Enterprise

Meta's 6,500-person Applied AI team is on the verge of revolt after engineers were drafted into data-labeling work

A Wired investigation published June 12 found that Meta's three-month-old Applied AI group, which houses roughly 6,500 engineers and product managers, is experiencing a serious internal breakdown. The unit was assembled to generate training data for Meta's AI models, producing coding puzzles and synthetic examples that teach AI agents to operate computers.

Employees told Wired they were moved into the group through a surprise email that gave them two options: join or quit. Workers began calling themselves "draftees." The crisis came into the open when someone hijacked a live internal presentation with an expletive-laden outburst directed at a senior Meta AI executive. Sources described the work as "literally the gulag." Over 1,600 Meta employees across the company signed a separate petition protesting keystroke monitoring tied to AI training data collection.

The episode surfaces a structural problem that will recur at any company scaling internal data operations: experienced engineers hired to build products resist being converted into annotation labor, and the resistance can become public quickly. Full coverage: Meta Applied AI revolt

SpaceX closes its IPO up 19% at $160.95, making Elon Musk the world's first trillionaire

SpaceX went public on the Nasdaq on June 12, pricing 555.6 million shares at $135 each to raise $75 billion in what the company described as the largest IPO in history. Shares opened at $150, an 11% pop from the offering price, and closed at $160.95 for a 19% first-day gain. The offering gave SpaceX a market cap above $600 billion.

At those prices, Elon Musk, who owns roughly 42% of the company, became the world's first trillionaire. SpaceX operates a subsidiary called SpaceXAI that provides satellite-enabled AI inference services. Musk's separately incorporated xAI (Grok) shares investors and strategic ties with his broader portfolio but did not participate in this offering.

The timing matters for the AI sector: Anthropic and OpenAI have both filed confidential IPO paperwork. SpaceX's strong first-day trading sets a high-profile reference point for how capital markets are pricing ambitious technology bets in mid-2026. Full coverage: SpaceX IPO and Musk trillionaire

Sources