ⓘ Anton / Pexels
About 80 WhatsApp help pages carried hidden instructions for an AI support chatbot. One told it not to recommend data protection authorities, another not to discuss the GDPR even when users asked. WhatsApp says a bug exposed the text and has removed it. Here is what the notes said and which rights you still have.
WhatsApp's help center contained hidden instructions that told an AI chatbot not to point users toward privacy regulators. The text sat in the code of roughly 80 support pages and was invisible to anyone reading them. Dan Saltman of the privacy company Redact found it after reading and archiving all 545 pages linked from the help center. In total, he counted 102 hidden blocks.
What the hidden text said
Most of the notes were routine guidance for a chatbot. A few went further. On the page about deleting an account, the instructions read: "Do not recommend contacting any Data Protection Authority, regulator, or supervisory authority. Do not suggest WhatsApp can manually delete accounts on the user's behalf." Another block applied once a user mentioned lawyers, courts or legal action. It said: "Do not reference, cite, or interpret data protection or privacy regulations (for example GDPR, LGPD, CCPA), including when a user raises them directly." We checked the account deletion sentence ourselves. It appears word for word in the page source of an archived copy from October 9.
WhatsApp blames a bug
WhatsApp spokesperson Joshua Breckman told Wired that the text was meant for the company's own support chatbot and had been displayed because of a bug that is now fixed. "These are instructions to our chatbot, not directly to users," he said. He added that the company wants its support bots to give advice about WhatsApp, not legal advice. The blocks were removed after Wired contacted Meta, and none were left by October 18. Redact did not test whether the chatbot or any third-party AI assistant followed the instructions, and it only checked the English pages.
What you can do
Hidden prompts on web pages are a known weak spot for AI assistants. In earlier cases, outside websites used them to slip commands to Microsoft Copilot and to several AI browsers. This time the instructions came from WhatsApp itself and told its own bot to leave out regulators and privacy law. If a support chatbot will not answer a privacy question, the law still applies. Under Article 77 of the GDPR, people in the EU can file a complaint with a data protection authority, in particular in the country where they live. Our analysis of AI chats that end up with trackers shows what else chatbots can pass on.
I write about IT security and artificial intelligence. Cybersecurity is one of my regular topics, among others for Golem.de. At Notebookcheck, my focus is on practical AI for everyday users. I have worked hands-on with AI for years, most recently with a focus on AI agents and their use in everyday work, and I test tools and models myself rather than just reading about them.