Anthropic Adds ID Checks, Lobbies DoD, Eyes EU Rules

Three separate regulatory and compliance moves this week show Anthropic working multiple fronts at once: tightening access to Claude, pursuing U.S. defense contracts through a new lobbying hire, and engaging European regulators over its cybersecurity AI models.

Key points

• Claude now requires government-issued ID and a selfie via Persona for identity verification, raising privacy questions about data handling and cross-border transfers.
• Anthropic hired Ballard Partners, a lobbying firm with close ties to the Trump administration, specifically for advocacy around “DOW procurement”, DOW being the Trump administration’s preferred name for the Department of Defense.
• Anthropic is in active discussions with European Commission officials about its cybersecurity-focused AI models and their regulatory treatment under EU frameworks.
• The Ballard Partners disclosure is the first confirmed external lobbying hire tied directly to Anthropic’s defense procurement ambitions.
• Persona’s involvement in Claude’s ID verification raises questions about third-party data access and the treatment of user data for those outside the United States, particularly in India.

What is the Persona verification rollout?

Anthropic has begun requiring users to verify their identity through a process involving a government-issued ID and a selfie, handled by the third-party identity verification provider Persona. The move is designed to strengthen access controls on Claude, but it has drawn scrutiny on two counts.

First, Persona is a U.S.-based vendor, which means biometric and identity data flows through an American company’s infrastructure. For users in India and other jurisdictions with strict data localization or cross-border transfer rules, that raises compliance questions that Anthropic has not yet addressed publicly. Second, the collection of selfies and government IDs represents a significant escalation in the personal data Anthropic holds on its users, beyond the typical email-and-password account model common across AI platforms.

The privacy concerns are notable from an investor perspective. As Anthropic pursues a path toward a public listing, any regulatory action over biometric data mishandling in the EU or India could become a material risk factor. The EU’s GDPR imposes strict rules on biometric data, and India’s Digital Personal Data Protection Act adds a further layer of complexity.

Why hire Ballard Partners now?

Ballard Partners filed a disclosure stating it was retained for advocacy specifically regarding “DOW procurement,” using the Trump administration’s terminology for the Department of Defense. The firm has well-documented relationships within the current administration, making it a logical choice for a company trying to move faster on federal contracts.

This hire follows a period of significant legal friction between Anthropic and the Pentagon. The company won a preliminary injunction earlier this year blocking a supply-chain restriction, and the DOJ subsequently filed an appeal. Hiring a connected lobbying firm while that litigation continues is a calculated move: Anthropic appears to be pursuing the defense relationship through both legal and political channels simultaneously.

For IPO watchers, government revenue is a meaningful part of the bull case for AI companies. Securing durable DoD contracts would add a recurring, large-contract revenue line that institutional investors value highly when underwriting pre-IPO AI companies.

What is happening with EU cybersecurity talks?

The European Commission has confirmed it is in ongoing regulatory discussions with Anthropic over its cybersecurity AI models. The specifics of what is being negotiated are not yet public, but the confirmation itself is notable. Most large AI developers have spent the past two years reacting to EU regulatory pressure rather than proactively engaging it.

Anthropic’s decision to open talks on cybersecurity AI specifically is likely tied to the EU’s broader work on AI and critical infrastructure rules, where cybersecurity applications occupy a higher-risk category. Getting ahead of those classifications, or shaping how they are drawn, would have commercial consequences across the European market.

The bigger picture

Taken together, these three moves point to a company that is deliberately building out its regulatory surface area ahead of a public offering. Identity verification controls speak to enterprise and government buyers who need audit trails. The Ballard Partners hire is a direct bet on federal revenue. The EU engagement is market access protection. None of these are product announcements; all of them affect the risk and revenue profile that underwriters will scrutinize when Anthropic’s IPO timeline firms up.