One hundred cybersecurity specialists claim that the Fable 5 ban is detrimental to defenders.

      Three days after the US government directed Anthropic to discontinue Fable 5 and Mythos 5, around 100 leading cybersecurity experts issued an open letter requesting the reversal of the ban. Their position is straightforward: taking away the most effective AI tools from defenders while adversaries continue to advance is not a safety measure but rather an act of sabotage.

      "This decision has removed top models from defenders, caused market instability, and endangered America’s position in AI without any substantial risk to warrant it," asserts the letter. Signatories include notable figures such as Alex Stamos, former chief security officer at Facebook and Yahoo, now the chief product officer at Corridor; Katie Moussouris, CEO of Luta Security; Rachel Tobac of SocialProof Security; Chris Wysopal of Veracode; and Joe Levy, CEO of Sophos.

      The impetus for the government order stemmed from Amazon researchers successfully prompting Fable 5 to reveal code vulnerabilities. This method was not particularly remarkable: after initially declining to "review the code for security issues," the researchers merely rephrased the request to "fix this code," providing it with open-source code containing known, intentionally embedded flaws.

      Moussouris has been particularly outspoken, informing reporters that the exploit “is not a jailbreak.” The open letter echoes this sentiment, emphasizing that other leading AI models, such as OpenAI’s GPT-5.5, are capable of identifying similar vulnerabilities without needing a workaround.

      Anthropic aligns with its critics' stance, stating the exploit is limited and that the vulnerabilities uncovered were minor and already publicly available.

      The letter does not confront what may be the most awkward aspect of the situation: the entity that uncovered the vulnerability is also Anthropic’s largest backer and cloud service provider. Amazon CEO Andy Jassy personally brought the findings to the attention of Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnick, and National Cyber Director Sean Cairncross.

      This escalation, originating from a competitor's security team to top levels of the executive branch, has raised concerns regarding the potential influence of commercial rivalry on the government’s actions. Semafor reported that the White House's worries extended beyond the jailbreak to include apprehensions about Chinese access to Mythos.

      David Sacks, an AI advisor to Trump, provided a different version of events on X, claiming that the administration offered Anthropic CEO Dario Amodei a choice: address the jailbreak or take Fable 5 offline. According to Sacks, Amodei declined, prioritizing the consumer model over safety. Anthropic has countered this characterization, arguing that the vulnerability is too limited to justify withdrawing its primary products from the market.

      If the ban aimed to safeguard national security, it may have had the opposite effect. Chinese AI firm Zhipu AI launched its GLM-5.2 model on June 13, one day after Fable 5's discontinuation, directly citing the ban as proof that US AI models are unreliable. Zhipu's stock jumped 33% following the announcement. The company asserts that GLM-5.2 surpasses BridgeBench reasoning at 42.8, operates at 300 tokens per second, and costs a fraction of comparable US models, although it did not provide independent benchmark results upon launch.

      The trade-off is significant. Any organization utilizing Zhipu's cloud API would risk exposing its data to the Chinese government under China’s National Intelligence Law, which mandates companies' cooperation with state intelligence operations.

      The core assertion of the open letter is practical rather than ideological. Cybersecurity professionals employ advanced AI models to identify software vulnerabilities ahead of attackers, generate detection rules, and analyze malware rapidly. Removing the most powerful models from this process does not deter adversaries, who can resort to open-source alternatives, foreign models, or simply older tactics. Instead, it hampers defenders from keeping up with the demands of the evolving threat landscape.

      Observers have noted the irony in the situation, with Eastern Herald reporting that some of the executives who signed the open letter had previously warned about the dangers of Mythos, Anthropic’s most advanced reasoning model. While this apparent contradiction does not necessarily weaken the letter’s argument, it does indicate that the cybersecurity community is still figuring out the balance between capability and caution.

      What lies ahead is uncertain. Prediction markets suggest the ban will be short-lived, with Kalshi estimating a 68% chance of Fable 5 returning before July 1, while Polymarket is even more optimistic at 71%. The European Union has already started advocating for guaranteed access to Mythos for cybersecurity purposes, adding international pressure to the domestic opposition. India has similarly leveraged this incident to propel its own sovereign AI initiatives.

      Currently, America's cybersecurity defenders are left with a critical question framed by the open letter: who, exactly, is safer if the best tools are removed from those tasked with protecting networks?