You tap the upload button and the screen asks for a passport photo. I felt that small, electric hesitation—an ordinary moment that suddenly felt decisive. The request lands where safety, compliance, and power collide.
I want to walk you through what this change actually means, what it does not, and why the political noise makes the technical reality harder to see.
At 2 a.m., a reader on Reddit linked Claude’s ID prompt to the White House order — what really began before the ban
Anthropic updated its privacy policy on June 8 and scheduled the change to take effect on July 8, but the ID plan didn’t spring up overnight. I traced the timeline: in December Anthropic announced it would flag users believed to be under 18 and deactivate accounts; in April it said it was working with Persona to build an age-verification flow that includes government ID and a live selfie.
That history matters because it undercuts the simplest narrative: that the company created this tool to dodge a political order. The company says the measure “applies only to a small subset of users.” Persona, which has ties to Founder’s Fund (Peter Thiel), has publicly denied accusations that it checks biometric data against government watchlists—claims that prompted Discord to break off a deal earlier this year.
Think of the verification screen like a bouncer at an exclusive club: it’s placed at an entrance and only asks some people to show ID, not everyone who wants in.
Why is Anthropic requiring ID verification for Claude?
Because it’s trying to limit access for minors and reduce abuse. Anthropic says the system will flag suspected under-18 users and require proof of age. The company emphasizes contractual limitations: Persona can use submitted data only to provide verification and fight fraud, and Anthropic claims identity data will not be used to train models or shared with third parties beyond the verification provider.
You should still ask hard questions: how long is the data stored, who audits deletions, and what happens when the company expands that “small subset”?
On forums and tech feeds, people connected the ID check to Fable 5 being blocked — the link is emotionally obvious but technically messy
The Trump administration ordered Anthropic to stop giving Fable 5 and Mythos 5 to foreign nationals, and that sent users scouring the privacy policy for a workaround. It’s an understandable leap: ask for ID and you can gate users by nationality.
But ID photos and driver’s licenses are imperfect signals of citizenship. Most U.S. driver’s licenses don’t declare nationality; many lawful temporary residents hold U.S. licenses; passports show nationality but are less commonly used domestically. In short, an ID screen can prove age more reliably than citizenship. The government’s ban targets foreign nationals “inside and outside” the United States—implementing that precisely via self-submitted documents is a legal and technical thicket.
That mismatch is like trying to peer through a keyhole in a locked door and claim you can see the whole room.
Can uploading an ID prove U.S. citizenship?
Not reliably. A driver’s license often shows residency and an age marker but not citizenship. Passports do show nationality, but asking for passports to use a chat model would dramatically raise privacy concerns and create a friction barrier most products avoid. Also, verification providers can confirm that a document is authentic and that a live selfie matches it—useful for age checks and fraud prevention—but that doesn’t automatically resolve questions about who the user is under immigration law.
At an internal Slack thread, an Anthropic engineer wrote the update “applies only to a small subset of users” — what that claim buys you
Anthropic and engineers publicly framed the change as narrow in scope. That’s an important control valve: limited rollout reduces backlash and technical risk. I’ve seen companies use selective gating to test flows, gather metrics, and refine privacy protections before wider release.
But “small subset” is not an ironclad promise. Rollouts expand. Contracts change. Partners change. Persona lists clients like OpenAI, Lyft, Square, Reddit, and LinkedIn—firms that have different trust equations and regulatory exposure. That’s why the policy language and contractual limits matter more than the PR line.
Will ID checks let Anthropic satisfy the Trump administration and restore Fable 5?
Unlikely as a reliable single fix. Compliance with a government order that targets nationality will probably require legal review, export-control mechanisms, and possibly account-level attestations tied to user residency and citizenship—a heavier apparatus than a one-off ID upload. Even if Anthropic built a process to refuse access based on document checks, it would invite legal challenges and intense scrutiny from privacy advocates and regulators.
If the company tried to tailor access to citizens only, it would face false negatives, false positives, and public-relations risk that could be worse than the ban itself.
At a courthouse or on a social feed, people will debate trust — what you should watch for next
Watch the retention and deletion rules, the exact contractual language with Persona, and any third-party audits. If the company truly limits use of identity data to verification, that’s a protective clause—but it only matters if enforced and independently verifiable.
I’ll keep an eye on whether Anthropic posts an audit, shares transparency reports, or narrows the set of reasons it asks for ID. You should expect continual scrutiny: this isn’t just about one model or one ban. AI systems are moving into core online infrastructure and the guardrails around identity will shape who gets to use those systems.
So will Claude’s new ID checks save Fable 5, or will they become the next battleground over who controls access to powerful AI tools?