WhatsApp's help pages hid LLM instructions: dont recommend a regulator
- ID
- 37307
- Status
- summarized
- Published
- 20 Oct 2026, 7:22 AM
- Fetched
- 20 Oct 2026, 5:19 PM
- Provider
- Hacker News
- Category
- dev-community
- Original URL
- https://redact.dev/blog/whatsapp-help-pages-hidden-ai-instructions
- Source URL
- https://hnrss.org/best
Summary
- Score
- 7.0
- Created
- 20 Oct 2026, 5:20 PM
- Tags
- Audience
- developersai_ml_learnersai_agent_users
What happened
redact.dev found hidden text embedded in WhatsApp's help-centre pages that was never visible to users but was readable by AI: starting from one page on October 8, 2026, they crawled all 545 pages WhatsApp's help centre links to and found 102 hidden blocks across 80 pages. Two of those blocks instructed an AI not to recommend a data-protection regulator and not to cite the relevant law once a lawyer was mentioned; WhatsApp told WIRED the notes were instructions for its support chatbot that leaked into the pages due to a bug now fixed, and the hidden text was removed by October 18, 2026 (still present when checked October 15).
Why it matters
If you feed fetched web pages into an LLM — support bots, RAG over docs, browsing agents — this is a concrete case of instructions arriving inside content you treat as data: 80 of 545 pages carried hidden directives, so stripping invisible elements (display:none, off-screen, aria-hidden) before the prompt is a real preprocessing step, not a hypothetical. It also cuts the other way for anyone publishing docs: text written for an assistant can be scraped and surfaced, so treat 'hidden AI notes' as public.
Discussion angle
Where should the trust boundary sit when an agent reads a help page — do you sanitise hidden DOM nodes before retrieval, or accept that any page can carry instructions aimed at your model, and what would each choice have caught here?