WhatsApp Hid Instructions in Help Pages Telling Its AI Chatbot Not to Recommend Privacy Regulators
evilsocket · x · 2026-10-20
Per The Hacker News and WIRED, data-deletion firm Redact found hidden instructions embedded in WhatsApp's help pages telling its AI support chatbot not to recommend contacting data protection authorities—even though WhatsApp's own European privacy policy tells users they can complain to one. The text was removed by October 18.
- The hidden block sat beneath the visible note on account deletion (deactivated accounts deleted after 30 days, inactive after 120, or 365 in the US), repeating the waiting periods and adding orders against suggesting regulators or legal routes
- A second hidden block instructed the bot not to cite privacy laws like GDPR once a user mentioned a lawyer, lawsuit, or court
- WhatsApp said the instructions were for its AI support bot, appeared due to a bug now fixed; whether the chatbot ever followed them is unknown
More from Safety
- Pivotal opens London research fellowship: 15 weeks, £15-18k stipend, mentors from Anthropic and DeepMind — davidmanheim · 2026-10-21
- New safety paper builds robust sandbagging model organisms for ELK research — ethayarajh · 2026-10-21
- Dormant poisoning can turn safe AI agents dangerous after benign fine-tuning, UNC team warns — mohitban47 · 2026-10-21
- Anthropic rolls out weight-signing on all internal token traffic per Frontier Safety Roadmap update — adam_karvonen · 2026-10-21
- Joshua Saxe: Curb AI cyber misuse without blocking open weights — joshua_saxe · 2026-10-21
- What rules would you need to trust an agent with write access? A builder's draft framework — Fresh_Spread_9223 · 2026-10-21