Anthropic’s apparent inability to identify which of its users are foreign nationals has led to some collateral damage from a US export ban on its most powerful AI models — but there is a way around it, at least for some.
On Friday, the US government ordered Anthropic to suspend access to Fable and Mythos, the new AI models it had introduced just a few days earlier, to all foreign nationals, citing national security reasons.
While the drafters of the US order may have had sovereignty in mind, they ended up making it an identity management problem. “The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” Anthropic said in a blog post commenting on the order, implying that it was unable to distinguish between foreign nationals and US citizens in its user base.
That’s likely the case today, but for its consumer customers, an update to its privacy policy, introduced last week and taking effect on July 8, gives it a new option: asking them for government ID.
The section of the policy on collection of personal data contains a new provision under the heading “Personal data you provide to us directly,” saying:
If the government ban on foreign access to Fable and Mythos continues, that would give Anthropic the option of opening access to users willing to submit a scan of their identity document, provided that it contained proof of their US citizenship. That would be the case for US passports — and also for citizens’ driving licenses issued by some US states along the country’s Northern border, which issue so-called enhanced driving licenses indicating the holders’ nationality. Enterprise users most likely to benefit from the power of the new AI models, though, will have to hope Anthropic finds some other way out of the current impasse.
The article originally appeared on CIO.