TL;DR
White House AI adviser David Sacks and Anthropic are offering conflicting accounts of why Anthropic’s Fable models were restricted. Sacks says a jailbreak could restore cyberweapon-like capability; Anthropic says the issue was narrow, minor, and reproducible in other models. The technical record remains non-public.
White House AI adviser David Sacks has given a detailed public account of why the administration restricted Anthropic’s Fable models, saying the company refused to fix a jailbreak that could restore Mythos-level cyber capability; Anthropic says the flaw was minor, already known, and not specific to its systems. The dispute matters because a major AI policy action is now being defended and challenged on evidence the public cannot review.
Sacks, who co-chairs the President’s Council of Advisors on Science and Technology, said on X that a “highly credible trusted partner” found a way around Fable’s guardrails. According to Sacks, the administration asked Anthropic to fix the issue or pull the model, and the company refused. He said the resulting export-control action was issued “reluctantly.”
Anthropic gave a sharply different account in a June 12 blog post. The company said the government had not provided specific technical details and described the demonstration it saw as involving minor, already-known flaws. Anthropic also said other public models, including GPT-5.5, could produce similar results without the same alleged bypass.
Reporting cited by Thorsten Meyer AI says Semafor, carried by Fortune and others, identified Amazon as the possible “trusted partner” that flagged the issue, with Amazon CEO Andy Jassy reportedly in contact with the administration. Amazon has not confirmed those specifics. That point matters because Amazon is an Anthropic investor and cloud provider while also competing in AI models.
The Safety Card, Played From Every Side
● ContestedA White House adviser says Anthropic refused to fix a cyberweapon jailbreak and got banned for it. Anthropic says the flaw is trivial. Almost every fact that would settle it is non-public — and “safety” is now the card every side is playing.
Both are claims, not findings. They don’t disagree on tone — they disagree on what the bypass actually is.
- A “highly credible trusted partner” found a jailbreak of Fable’s guardrails.
- The admin asked Amodei to fix it or pull the model. He refused.
- So the export control was issued — “reluctantly.”
- It restores operability of a cyberweapon; calling that “not serious” is indefensible.
- The government gave no specific technical detail.
- The demo found a few minor, already-known flaws.
- Other public models (incl. GPT-5.5) do the same without a bypass.
- A “narrow potential jailbreak” shouldn’t recall a model used by hundreds of millions.
Per reporting by Semafor (carried by Fortune and others), the entity that flagged the jailbreak was Amazon — with CEO Andy Jassy reportedly in contact with the administration. Amazon hasn’t confirmed specifics. Flagging a real risk is what a good partner does — but Amazon wears three hats at once, and none of them is neutral.
Each actor’s safety claim points toward its own advantage.
The entire evidentiary record is a matter of trusting parties who each have a reason to shade it.
A transparent, technically grounded, independently reviewable process — which is, notably, exactly what Anthropic says it wants, and exactly what would also constrain Anthropic. The reason to demand it isn’t loyalty to anyone; it’s that the alternative is decisions made on secret evidence and adjudicated in dueling press statements.
Independent commentary, produced with AI assistance under human editorial oversight; the views are the author’s own and may change. This is analysis and opinion, not investment, financial, legal, or technical advice, and it concerns an actively developing situation in which key facts are disputed and non-public. Claims attributed to David Sacks reflect his June 13, 2026 statement on X; claims attributed to Anthropic reflect its published statements; reporting on Amazon’s role reflects accounts published by Semafor and others — all read as of June 15, 2026, and presented as the claims of those parties, not as established fact. Characterizations are the author’s interpretation, offered in good faith and open to rebuttal. References to specific people, companies, and government actions are factual and analytical, not partisan, and imply no affiliation or endorsement.
Secret Evidence Tests AI Oversight
The case has become a test of how governments should act when AI safety claims involve powerful models, classified or private technical evidence, and commercial rivals. If Sacks’s account is accurate, the administration acted to limit access to a model that could be used for advanced cyber operations. If Anthropic’s account is closer to the facts, the action may have punished a company over a narrow issue that independent reviewers have not seen.
The dispute also cuts against simple alliances. Anthropic has often argued for stronger safety rules around frontier AI, including models with advanced cyber capability. In this case, the same safety logic is being used against Anthropic. That makes the process behind the decision as consequential as the technical flaw itself.
Ai Engineering Made Practical: Build Reliable Ai Systems With Retrieval, Tools, Evaluation, Monitoring, And Safety—So Teams Ship Faster With Less Risk
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Fable Follows Mythos Warnings
The dispute centers on Anthropic’s Fable model line and its relationship to Mythos, a more powerful system Anthropic had previously framed as carrying serious cyber risk. Sacks’s account says Fable is effectively Mythos with guardrails, meaning a successful guardrail bypass could expose dangerous capability. Anthropic disputes the severity of the alleged bypass.
The source material describes the episode as the most detailed government-side explanation yet for the action against Anthropic’s models. It also stresses that both sides are making claims rather than presenting a full, reviewable record. No public CVE, technical write-up, named testing lab, or independent assessment has been released.
“A “highly credible trusted partner” found a jailbreak of Fable’s guardrails.”
— David Sacks, via X
AI model jailbreak detection software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Missing Technical Record
It is not yet clear what the jailbreak does, how it was tested, whether it is unique to Fable, or whether it reliably enables the cyber capability Sacks described. The identity and role of the “trusted partner” also remain unconfirmed by the parties involved.
The public record does not show whether Anthropic refused a specific patch request, disputed the technical finding, or objected to the process. It is also unclear whether any independent evaluator has reviewed the alleged exploit.
Volam 4G LTE Cellular Dash Cam with Live Streaming & GPS Tracking – Dual Car Security Camera with Remote Live View,Cloud Storage,AI Motion Detection,Sentry Mode, Anti-Theft Protection for Vandalism
[4G LTE CELLULAR CONNECTIVITY & LIVE STREAMING FROM ANYWHERE] Stay connected to your vehicle 24/7 with our advanced…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Patch Timing May Signal Severity
The next signal will be whether the restriction is lifted quickly after a quiet fix or remains in place while the parties continue to dispute the evidence. A fast reversal could support the view that a concrete patch was available. A longer standoff could strengthen Anthropic’s argument that the issue was not the severe model-specific failure Sacks described.
Outside experts and competitors are likely to press for a reviewable process for future model restrictions. That could include technical methodology, independent testing, and clearer rules for how safety reports from commercial partners are handled.
Safetec Universal Precaution Compliance Kit (Poly Bag)
Safetec
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What happened to Anthropic’s Fable model?
A U.S. restriction was placed on Anthropic’s most powerful models, according to the source material. David Sacks says the action followed Anthropic’s refusal to fix or pull Fable after a jailbreak was reported. Anthropic disputes that account.
What does David Sacks claim?
Sacks says a trusted partner found a guardrail bypass that could restore Mythos-class cyber capability in Fable. He says the administration asked Anthropic to address it and acted only after the company refused.
What does Anthropic dispute?
Anthropic says the government did not provide specific technical detail and that the demonstrated flaws were minor and already known. The company also says similar outputs can be produced by other public models.
What role may Amazon have played?
Reporting cited in the source material says Amazon may have been the partner that flagged the issue, though Amazon has not confirmed the specifics. Amazon’s possible role draws attention because it has business ties to Anthropic and also competes in AI.
Why does the lack of public evidence matter?
Without a technical record, outside observers cannot judge whether the restriction was a necessary safety measure or an overbroad action based on disputed claims. That gap leaves the decision resting on trust in parties with policy, business, or competitive interests.
Source: Thorsten Meyer AI
