WhatsApp Bug Exposed Internal Instructions for Meta Chatbot
· Technology · Wired, India Today
Dozens of WhatsApp help and FAQ pages inadvertently exposed internal text instructions designed for Meta's large language model chatbots. Privacy firm Redact discovered approximately 80 webpages containing hidden guidelines that instructed the AI on how to handle user queries. The exposed code included strict directives labeled “WHAT NOT TO SAY,” which specifically prohibited the chatbot from providing legal advice or recommending data protection authorities. Meta removed the text from all affected pages after being contacted by media outlets, attributing the exposure to a technical bug. WhatsApp spokesperson Joshua Breckman confirmed that the instructions were intended solely to ensure accurate automated support responses and were not meant for public viewing.
Why it matters
The accidental disclosure reveals the specific guardrails and behavioral constraints tech companies implement when deploying AI assistants for customer support.
Read the original report — Wired
Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.