**Nathaniel Whittemore** (0:00)
In this emergency episode, we are discussing the US government shutting down Anthropic's Fable 5
The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.
All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, Robots and Pencils, Scrunge, Blitzy, and OutSystems. To get an ad-free version of the show, go to patreon.com/aidailybrief, or you can subscribe on Apple Podcasts. If you want to learn more about sponsoring the show, send us a note at sponsors at aidailybrief.ai. And while you're there, check out the new site. If there is any specific part of a specific episode that you want to share with someone, whether it's a number or some stat or quote, there's a good chance that it is now there, cut up and shareable for you, so go check it out, aidailybrief.ai.
Hello friends, welcome back to the AI Daily Brief. For the first time in the three-year history of this show, news has broken on a Friday afternoon that is too significant to wait until Monday to explore. Last night, just before 9 Eastern time in the US., Anthropic tweeted, The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national. Whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other cloud models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. After this absolutely stunning news, journalists and internet sleuths flew into a tizzy to try to figure out what the heck had actually just happened. The Wall Street Journal added some color, reporting that Commerce Secretary Howard Lutnick had sent a letter to Anthropic CEO Dario Amade announcing that the new models, Fable 5 and Mythos 5, were now subject to export restrictions, meaning usage by customers outside the US as well as foreign nationals within the US would be prohibited. So where did this seemingly capricious policy come from? It was apparently a report from another company about a jailbreak it had discovered. Anthropic gave more details in their blog post writing, We received the directive from the government today at 521 PM. The letter did not provide specific details of its national security concern. Our understanding is that the government believes that it has become aware of a method of bypassing or jailbreaking Fable 5 We reviewed a demonstration of this specific technique being used to identify a small number of previously known minor vulnerabilities. These vulnerabilities all appear relatively simple and we have found that other publicly available methods are able to discover them as well without requiring a bypass.
And basically from there went on to say that they just don't buy the U.S.'s logic. They point out that in the weeks leading up to the release of Fable, they worked with the US government and many others to read team Fable safeguards for a significant amount of time. They pointed out that, quote, no testers have yet been able to find a universal jailbreak, a jailbreak method that can very broadly bypass the model safeguards. Indeed, they write, we suspect that perfect jailbreak resistance is not currently possible for any model provider. Every safeguard used in the industry is vulnerable to non-universal jailbreaks, which can elicit some cyber information in specific circumstances, and it is likely that universal jailbreaks will eventually be found in the future. They said, given that perfect jailbreak resistance does not appear to be possible today, Anthropic adopted a defense-in-depth strategy with Fable 5 We aimed to make jailbreaks either narrow or very expensive to produce, and to combine this with thorough monitoring to quickly detect and shut down any successful attacks. Harkening back to the controversy of this week, they continue, This is also why Anthropic has required 30-day retention of customer data with Fable, a policy change that carries real costs for us with customers, but that allows us to research and mitigate jailbreaks.
Importantly, they conclude, we have not even received a disclosure of a concerning non-universal potential jailbreak that led to a harmful result. The potential jailbreaks that have been disclosed to us are either entirely benign responses or are minor findings that provide no mythos-specific uplift. To date, they write, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws. Our understanding is that one potential jailbreak was shared with the government. We have reviewed a report that we believe is the basis of the government's directive and validated that the level of capability displayed there is widely available from other models including OpenAI's GPT-55 and is used every day by the defenders who keep systems safe. Given that then, they write, we are complying with the government's legal directive and moving access to Fable 5 and Mythos 5 for all users. However, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers. Now, the Wall Street Journal later added, the jailbreak research in question was done by researchers at Amazon, who used a series of prompts to get Anthropics models to provide them with information about a handful of security vulnerabilities.
24 more minutes of transcript below
Try it now — copy, paste, done:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090
Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000772552056