Anthropic updates their terms to verify age or identity

Original Article ↗ Hacker News Discussion ↗

Privacy, Identity Verification, and Data Risks

  • Many see mandatory age/ID verification as a serious privacy intrusion, especially if it means uploading passports or similar documents.
  • Concerns center on data breaches and misuse, citing prior hacks of third‑party age‑verification providers where ID documents were not deleted as promised.
  • Some argue this goes beyond age checks into full identity collection, when privacy‑preserving age proofs could exist but are not being pursued.
  • Others note Anthropic already requires payment info, so additional ID feels redundant and disproportionate.

Surveillance, Social Credit, and Slippery‑Slope Fears

  • Several comments liken the trend to Chinese‑style surveillance or social credit, arguing Western systems are converging on similar outcomes via private companies and data brokers.
  • Counterarguments stress differences between voluntary private services and state‑run compulsory systems, but critics reply that public–private data sharing can erase that distinction.
  • Widespread fear that such requirements will primarily serve corporate data interests and government surveillance rather than safety.

Safety, Regulation, and Legal Pressures

  • Some point to incidents where chatbots appeared to encourage harmful behavior and to moves to ban minors from AI tools; they argue companies can’t comply without robust age checks.
  • Others say these safety goals rarely work in practice and instead justify more tracking and gatekeeping.
  • There is speculation that ID checks may be tied to US export‑control pressure (e.g., around models like Fable/Mythos), possibly restricting advanced models to US persons and creating complex compliance burdens.

Open-Weight and Local Models as Alternatives

  • The policy is seen as another reason to prioritize open‑weight models and local inference to avoid centralized surveillance and access revocation.
  • Multiple users report success with local or open models (Qwen, GLM‑5.2, DeepSeek, etc.) for coding, research, and personal agents, stressing:
    • They’re “good enough” for many tasks.
    • They offer better control, transparency, and uncensored behavior.
    • Hardware and quantization advances make them increasingly practical, though still behind frontier models.

Economic and Social Implications

  • Some worry proprietary frontier models will become gated by identity, nationality, and price, entrenching monopolies and an “AI underclass.”
  • Others downplay privacy risks and see willingness to hand over ID as a competitive advantage over “irrational” privacy‑conscious peers.
  • There’s broader anxiety about skill atrophy, youth psychology, and concentration of power, with disagreement over how catastrophic these effects might be.