Britain's most tattooed man says UK's age check told him to "remove his face" (www.dexerto.com)

When Britain rolled out stricter age-verification rules for adult sites that require a live facial scan, tattoo celebrity Matthew Whelan — aka King of Ink Land — says the system repeatedly told him to “remove my face,” blocking access because it flagged his heavily tattooed visage as a mask. Whelan, who’s spent more than 1,600 hours getting inked, was denied entry to a webcam site after the filter misclassified his facial tattoos as an occlusion. The episode has become a viral example of the rollout’s rough edges and has renewed criticism of the policy’s usability and privacy trade-offs. For the AI/ML community this is a textbook problem of biased models and brittle heuristics: mask/occlusion detectors and liveness checks trained on narrow datasets can produce false negatives for non‑normative appearances (tattoos, scars, makeup) and discriminate against subpopulations. It highlights risks around dataset representativeness, evaluation gaps (false rejection vs. spoofing tolerance), and the tension between anti-spoofing measures and accessibility. Users are already seeking workarounds (VPNs, synthetic avatars), which underlines that brittle biometric gates can drive insecure behavior. Practically, this calls for broader datasets, transparent failure modes, alternative age proofs (document checks, cryptographic age tokens), and careful policy-design to avoid denying services to legitimate users while preserving privacy and security.