That escalated quickly...

My experience of Fable 5 this week was limited, and I didn't really notice much difference between it and the Opus models. I've been following people's experiences on Reddit, however, and it's been game changing for them. Simon Willison, also gave a hint of how 'proactive' it can be.

At the Thinking Digital conference last month, I attended a workshop which used a futures technique to imagine the US and China as AI superpowers in the same way we have become used to thinking of 'mobile' as being either Apple or Google. We thought through the regulatory implications of that, and the downstream impacts on technology, society, and democracy.

I think this is just the beginning.

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 Anthropic models will not be affected.

We received the directive from the government today at 5:21pm (ET). The letter did not provide specific details of its national security concern. Our understanding is that the government believes 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 models are able to discover them as well without requiring a bypass.

[...]

We are complying with the government’s legal directive and are removing 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.

As we have stated publicly, we believe the government should have the ability to block unsafe deployments, as part of a statutory process that is transparent, fair, clear, and grounded in technical facts. This action does not adhere to those principles.

---

Source: www.anthropic.com
Are.na block: ↗
Collection: ai-critique-policy