Claude safety guardrails block Latin digitization requests
Illustrious-Bass9651 · reddit · 2026-08-31
A Doctor of History & Theology reports frequent blocks by Claude's safety guardrails while digitizing Latin and Ancient Greek works. The content consists of publicly available papers and out-of-copyright materials. The user has filed a complaint with Anthropic without response and questions if data siphoning is involved.
More from Safety
- Guardian: AI should help US politicians listen, not just persuade — nordicinst · 2026-08-31
- Autonomous AI Viruses Will Be a Manageable Nuisance, Not an Existential Threat — binarybits · 2026-08-31
- Support bot went off-brand and we couldn't stop it in real time — seeking guardrail setups — Strong-Income-5925 · 2026-08-31
- curl's Daniel Stenberg Goes Public Over a Disputed CVE — theanonymousone · 2026-08-31
- What actually has the authority to stop your agent when it goes wrong? — No_Progress92 · 2026-08-31
- AI-driven cyber risk is top concern for financial stability — talkingatoms · 2026-08-31