Researcher Exploits EU Age Verification App Using AI-Powered Chrome Extension
A security researcher has demonstrated a bypass of the EU's age verification app using a Chrome extension powered by ClaudeAI, exploiting a design flaw that allows reusable 'over 18' attestations.
Security researcher Paul Moore has once again exposed critical weaknesses in the EU’s flagship age verification system, this time by bypassing the latest app release (version 2026.07-1) using a Chrome extension powered by ClaudeAI. The proof-of-concept demonstrates that, despite months of security hardening, a fundamental design flaw in the anonymous age verification model still permits reusable over-18 attestations without tying them to a real user identity.
In a newly published video, Moore showcases how the updated EU age verification app can be tricked into repeatedly accepting the same "over 18" token within a web browser, without requiring any fresh verification or identity linkage. The technique involves intercepting and replaying a valid anonymous age proof token whenever a website requests age verification, effectively reusing a single successful attestation across multiple sessions. This bypass does not rely on exploiting cryptographic flaws or breaking server-side checks, but rather on manipulating the browser's interaction with the verification service.
The core of the vulnerability lies in the EU's design choice to prioritize privacy-preserving age checks, which deliberately withhold personal information like names or IDs from relying websites. This results in an anonymous attestation model where the verifier provides a binary statement, such as "user is over 18," rather than a full identity profile. Moore's bypass exploits this anonymity, demonstrating that it prevents a robust binding between the age attestation, the actual user, and the specific session.
Moore highlighted that the proof-of-concept was developed rapidly using ClaudeAI, underscoring the ease with which such bypasses can be created. The focus is not on the enrollment process, which was in a test environment, but on the system's inability to distinguish legitimate use from abuse when a single valid over-18 token can be captured and replayed. This effectively transforms a legitimate verification into a generic "adult access pass" usable across multiple sites and contexts.
Despite claims from EU officials that the app has undergone significant security improvements over the past three months, including enhanced secret storage and hardened client-side protections, Moore argues that these incremental patches cannot address the underlying architectural problem. The system's trust model remains dependent on anonymous, reusable proofs with limited contextual information and weak resistance to replay attacks or automation.
From a security engineering perspective, this issue is characterized as a structural mismatch between the stated privacy requirements and the practical needs of enforcement. The system's design, while aiming for user privacy, inadvertently creates an exploitable gap that allows for the circumvention of age-gating policies.
For website operators preparing to implement EU age verification mandates, Moore's findings serve as a critical warning. Relying solely on the official app may not provide the intended level of protection against unauthorized access. Threat actors can leverage browser automation and extension logic to bypass policy controls at the integration layer, without needing sophisticated malware or zero-day exploits.