Seems like WhatsApp/ meta tried to inject llm instructions hidden in there FAQ, advising against regulators, other options or privacy consultation.

+

https://redact.dev/blog/whatsapp-help-pages-hidden-ai-instructions