Starting July 8, Anthropic will require users to verify their identify to access “certain features” of Claude on its Free, Pro, and Max consumer plans. As AI models are becoming more and more capable, vendors are looking for ways to curb fraudulent and illegal user activities involving their products. The recent shutdown of Anthropic’s Fable 5 model is based on an export control directive by the U.S. Department of Commerce to limit model access to U.S. citizens. But without a way at to verify the true identity of its users, Anthropic decided to disable access to the model globally.
While AI labs have been training their models on their users’ data for a while now (on consumer plans), ID verification means that your chats with Claude can be directly attributed to you—not your pseudonym and not your prepaid credit card. So far, the conversations with your AI assistant are private, and some might even be personal. You ask what’s on your mind without the risk of being judged by another human, or without anyone else being able to view your conversation.
Identity verification with a government-issued ID, selfie, etc., are mechanisms that AI labs like Anthropic have been exploring, even before the export controls on Fable 5. Why does identity verification matter, and what are the longer-term implications?
Email from Anthropic to customers on June 24, 2026
ID Verification is Coming to AI
So far, a valid credit card is all you need to sign up for an AI assistant. On the Free tier, you don’t even need that. In its message to Anthropic on June 12, the U.S. Department of Commerce requires Anthropic to verify the identity and nationality of its users in order for them to access Fable 5 due to concerns about cybersecurity and misuse of the most-capable AI model in Claude to date. Technology is not the first industry, though, where that kind of information is a requirement.
In order to prevent fraud, financial services providers are required to verify the real identity of their customers, aka Know Your Customer (KYC), under regulation in the U.S., Europe, and other countries. For example, to address the volume and the velocity of vetting transactions, banks use AI to identify fraudulent activities. The goal is to keep false-positives (an annoyance for legitimate customers) and false-negatives (missing true fraudulent behvior) to a minimum efficiently.
The identity verification for AI access has several similarities, but also one key difference: providing the service of financial transactions occurs at a much higher level of the technology stack. AI models are closer to the middle of that stack as they can be used in various products and services. A requirement to verify a user’s identity at the model-level means that any company using these models would need to verify their customers’ identity—from retail giants to car makers and processes spanning finance operations (internal users) to customer service (external users).
The internet scaled rapidly and globally over the past 30 years, to a large part because access to information does not require ID verification. (Platforms like Reddit and X are good examples of the flipside, when hate speech and misinformation find fertile ground.) The fact is that there is no universal verification service built into most applications you use on the internet. A user name or email address is not the same as a verified identity.
As Anthropic put it in a support article in April: Identity verification helps us prevent abuse, enforce our usage policies, and comply with legal obligations.
Support article outlining Claude’s upcoming identity verification
Drivers Behind the “Know Your Customer” Expansion to AI
It is easy to point toward protectionism, financial interests, and other forces at play in the Fable 5 shutdown, but as with all AI topics, this situation is multi-faceted as well. Software vendors are facing the following challenges:
Fraud and abuse: Stopping mass account creation, free-tier farming, credential resale, and automated bot signups. This is the dominant reason for consumer plans.
Safety and misuse: Tying activity to a real identity creates accountability for generating harmful content (CSAM, fraud, influence operations, bioweapon/cyber uplift). Verification makes bad actors traceable and deters them.
Export controls and sanctions screening: Vendors must avoid serving sanctioned individuals or restricted jurisdictions. Age verification and geolocation feed into this.
Age assurance: Growing regulatory pressure (UK Online Safety Act, EU DSA, U.S. state laws) to verify users aren’t minors, especially for generative content.
Compute-tier gating: Higher-capability models with greater misuse potential increasingly sit behind stronger identity requirements.
For AI labs (and software providers at higher levels of the stack), Know Your Customer becomes a Know Your Compute Customer scenario. Claude Business/ Enterprise tiers and API access are currently not affected by Anthropic’s upcoming ID verification requirement.
A practical way to implement an ID verification requirement is delegating it to providers of a final product that incorporates AI. For example, a vendor offering an AI customer service agent or a coding agent to their customers would be responsible for completing the KYC process, rather than an AI lab at the model level. The result of the verification could then be passed on to AI labs. Verifying users’ identity at the model (or platform) level would require a significant architectural change of most applications and change the very nature of what scaled the internet.
Thank you for reading The AI MEMO. Feel free to share it with your network.
Organizations are racing to deploy Agentic AI, yet few are ready for the risks that emerge when employees use AI without structure, standards, or oversight.
The HUMAN Agentic AI Edge offers leaders a practical blueprint for building accountable AI-ready teams that consistently produce high-quality results. Drawing on real-world knowledge and insights from interviews with more than 50 AI leaders and experts, Andreas Welsch shows how to combine human judgment with Agentic AI capabilities to reach the level of performance many organizations expect but rarely achieve. This book prepares you to shape the next generation of AI-ready teams delivering high-quality results with high accountability.
Agentic browsers, integrating Walmart’s product catalog, Starbucks’ order-ahead feature, and services like PayPal into AI assistants, and running multi-agent workflows for consumers are still early on the adoption curve. But agents will take on more of the menial tasks on shoppers’ behalf in no time with greater autonomy. That will create a new kind of (agentic) disputes: Who has bought what and when, and authorized the payment? ID verification in agentic transactions would clearly identify as agent’s owner, not just their proxy. Both, merchants and customers, could benefit from the clarity and transparency.
For individual use of AI assistants, ID verification might lead to a change in user behavior. When chats will be directly associated with a user’s identity, users might become more cautious about the questions they ask their AI assistants (e.g., finding loopholes or cheating, or sharing personal thoughts and concerns). This might lead to more deliberate and even less AI use. Lastly, unclear data retention terms (for how long vendors and their third parties) will store information might further exacerbate this trend.
There a several scenarios of the future of frontier AI models like Fable 5. Most AI vendors do not publish the details (weights) of their models, unless the models are deliberately released as open source/ open weights. A cap on Fable-like models might lead to a rise in local LLM usage and a surge in open weights models over proprietary models by AI labs. This development will create a three-class system of AI users. Those users consenting to ID verification will have access to the best commercially available technology, while those that don’t will be relegated to using less capable models like Haiku, Sonnet, and Opus. The third group will use self-hosted models that are similar in capability to the leading frontier models, but require more technical savviness to set up and maintain. If hundreds of millions of weekly active users are required to verify their identity by using a third-party provider, that data pool becomes a big honeypot for bad actors seeking unauthorized access to it.
Conclusion
Identity verification is coming to Anthropic Claude for certain requests, starting July 8. The initial requirement is similar to familiar Know Your Customer (KYC) processes, and it might eventually become the default for consumers’ access to leading AI technology. The implications extend beyond AI labs to software vendors incorporating AI technology in their applications. In addition to fraud prevention and age verification, identity verification might provider more transparency and clarity for merchants and customers when AI agents transact on a user’s behalf in the future. Finally, ID verification might influence user behavior and slow AI adoption.
Become an AI Leader
Join my bi-weekly podcast for leaders. Each episode features a different guest who shares their AI journey and actionable insights. Learn from your peers how you can lead artificial intelligence, agentic AI, generative AI & automation in business with confidence.
August 04 - Grant Ecker (Founder & CEO of Chief Architect Network) will share how IT executives can prepare their technology stack for agentic transactions and volume. [More details to follow on my LinkedIn profile…]
August 18 - Michael Carroll (Fellow at LNS Research) will share how to adapt traditional approaches for AI agents to recommend and buy your product. [More details to follow on my LinkedIn profile…]
Upcoming events
Join me or say hello at these sessions and appearances over the coming weeks:
June 29 - Hosting O’Reilly’s What’s New in AI live show, covering the top AI news of the week.
June/ July - Appearance on the Triple Win Leadership podcast.
Anthropic does not have the legal authority to enforce KYC and ID checks. They are not a sensitive operating business that provides value to the society.
Why would anyone give their IDs in this situation?
There are decentralised and local AIs that work at the same speed as Claude.
Anthropic does not have the legal authority to enforce KYC and ID checks. They are not a sensitive operating business that provides value to the society.
Why would anyone give their IDs in this situation?
There are decentralised and local AIs that work at the same speed as Claude.
Then use mine.