TechCapsules

AI-summarised brief · reviewed before publication

WhatsApp’s help pages contained hidden code instructing its AI support chatbot to refuse recommending privacy regulators or citing GDPR rights when users could not delete accounts. The instructions also barred the bot from referencing legal terms after a user mentioned litigation. Redact, a data‑deletion firm, discovered 102 such blocks across 80 pages, removed them by October 18, and confirmed the bug was fixed. The hidden text contradicted WhatsApp’s European privacy policy, which explicitly allows complaints to data protection authorities.