Anthropic crackdown by Trump administration: who benefits
Anthropic pulled Fable 5 and Mythos 5 after a Trump administration export control order, prompting debate over winners and losers.
TL;DR
- 01Anthropic pulled Fable 5 and Mythos 5 after a Trump administration export control order, prompting debate over winners and losers.
- 02The move forced Anthropic to take both models offline because the company said it could not reliably ensure users were not foreign nationals.
- 03Independent security experts cited in the discussion argued the same jailbreaks could have been found in other AI models, and they have signed an open letter asking the president to revoke the order.
Anthropic pulled its two newest models, Fable 5 and Mythos 5, after the White House issued an export control order that cited "national security concerns" and required the company to prevent foreign nationals from using those models. The move forced Anthropic to take both models offline because the company said it could not reliably ensure users were not foreign nationals.
What happened and how did this escalate?
Anthropic removed Fable 5 and Mythos 5 following a White House export control letter that cited "national security concerns," and the company stopped the models because it could not guarantee they would not be used by foreign nationals. Reporting and the episode discussion say the White House action followed concerns raised by Amazon researchers who allegedly found a way to bypass Fable 5's guardrails, and Amazon CEO Andy Jassy reportedly raised those concerns with the White House.
The shutdown unfolded quickly, described in the conversation as moving "really fast," and Anthropic opted for a blanket takedown because many of its employees are foreign nationals and the order lacked public specifics. Independent security experts cited in the discussion argued the same jailbreaks could have been found in other AI models, and they have signed an open letter asking the president to revoke the order. The experts warn that removing these models also pulls advanced cybersecurity capabilities from U.S. network defenders.
Who stands to gain or lose?
The immediate losers are users and defenders who had access to Mythos 5 and Fable 5, and Anthropic itself, which is now navigating both technical and political fallout. The episode panel noted that Anthropic's relationship with the Trump administration has been strained, and one host argued that tension may have made Anthropic a target in a way that could spare other labs.
Some analysts and security researchers framed the action as potentially retaliatory, citing a broader supply chain dispute and an ongoing lawsuit between Anthropic and the administration. Others pointed out that the technical risks cited are not unique to Anthropic, which raises the prospect that competitors gain a short-term advantage. The panel also flagged a reputational effect: prior clashes between Anthropic and the administration led to increased downloads of Claude, suggesting regulatory conflict can drive user interest. As one contributor put it, "everybody loves a bad boy." That public curiosity could paradoxically benefit Anthropic even as the company fights to restore access.
Why it matters
The order exposes how political friction can become a de facto regulatory tool and reshape which AI systems stay available. It also highlights a trade-off between restricting potentially risky models and removing tools that some defenders use to secure networks. Independent security experts who signed the open letter argue the export control approach may hamper U.S. cybersecurity capabilities, and the episode discussion stressed that perceived uneven treatment across labs introduces unpredictability into the AI market. The outcome will influence who feels safe releasing advanced models and whether companies will change how they talk about risk versus capability.
What to watch
Watch whether the White House responds to the open letter and whether the export control order is revoked. Track whether Anthropic restores access to Fable 5 and Mythos 5 and how competitors react: a repeat of the prior clash reportedly drove downloads of Claude, and similar shifts in usage or availability would confirm market effects described in the discussion.
Sources: the company action to pull Fable 5 and Mythos 5, the White House letter citing "national security concerns," the role of Amazon researchers and Andy Jassy raising issues with the White House, cybersecurity experts signing an open letter asking for revocation, and the panel's note that downloads of Claude increased after a previous blow-up are all from the cited episode conversation on June 18.
- a week before Fable came outAnthropic releases Fable 5 publicity
Anthropic had publicly positioned some models as powerful and in need of caution shortly before Fable's release.
- last FridayWhite House issues export control letter
The administration cited "national security concerns" and told Anthropic to ensure the models could not be used by foreign nationals, prompting Anthropic to take Fable 5 and Mythos 5 offline.
- June 18Experts and commentators react
Cybersecurity experts signed an open letter asking the president to revoke the order; commentators noted Amazon researchers had allegedly found a guardrail bypass and that prior clashes had boosted Claude downloads.
Written by The Brieftide · Source: TechCrunch
The Brieftide Daily · 06:00
Briefs like this one, in your inbox every morning.
Continue reading
More in AI SafetyAI4SE and SE4AI: A decade review of AI in systems engineering
H. Sinan Bank, Daniel R. Herber and Thomas Bradley map three research phases and assess 1.
Deepmind AI Control Roadmap: agents treated as insider threats
Deepmind ties permissions to verified behavior, models agents as rogue employees.
Dario Amodei's AI playbook: Anthropic's regulation plan
Amodei urges binding third-party audits, federal power to block risky models, export controls.
Germany approves DE-AISI, an AI security institute based on UK
The National Security Council authorised a German AI Security Institute to test advanced models.