Around 80 help and FAQ pages on WhatsApp's website exposed hidden instructions written for one of Meta's AI chatbots, Wired reports, citing an analysis by the privacy company Redact.
The text sat in the pages' code rather than on screen, and included sections headed "WHAT NOT TO SAY" and directions not to give legal advice. On the page about deleting an account, the bot was told not to recommend contacting data protection authorities or regulators.
Redact's chief executive, Dan Saltman, said the gap between what users are told and what chatbots are told stood out, though many instructions simply pointed people to official support channels. After Wired contacted Meta, the text was removed. A WhatsApp spokesperson said a bug had caused it to appear and that the instructions were meant to keep the support chatbot's answers accurate.
Source: Wired
