Anthropic:
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.
Fable 5 and Mythos 5 were announced on Tuesday — that post has detailed comparisons, from Anthropic, on the models’ capabilities.
Having the access ban extend all the way to “foreign national Anthropic employees” is, to say the least, aggressive. Whether that degree of restriction is truly warranted, I don’t think we, on the outside, can say. The Trump administration lacks credibility, to say the least, when it comes to foreign nationals. But it’s Anthropic itself that repeatedly compares the power of frontier models to nuclear weapons. Here’s CEO Dario Amodei, in an essay published just this month:
There may come a time, perhaps relatively soon, when we need to go
beyond this, when the most powerful AI systems look less like
airplanes or automobiles and more like weaponizable nuclear
materials — a threat to humanity rather than “just” a threat to
public safety. If that occurs, we may need more aggressive
regulatory measures than those I have laid out.
If that occurs — or if it already has occurred — it’s obviously not the place of Anthropic (or OpenAI or Google) to render that judgment. Ben Thompson wrote about this presciently back in March, and linking to his post, I wrote:
Nilay Patel, quoting the same section of Thompson’s column I
quoted above, sees it as “Ben Thompson making a
full-throated case for fascism”. I see it as the case against
corporatocracy. Who sets our defense policies? Our
democratically elected leaders, or the CEOs of corporate defense
contractors?
