WitchdoctorTech
Meta’s AI app, used by over a billion people, is leaking private conversations—including medical, legal, and deeply personal info—into the public Discover feed. Many users don’t realise that tapping the “Share” button makes chats visible to anyone, even without logging in.
From teachers disputing job terminations to folks asking about tax evasion, Meta AI has become a privacy minefield. Worse, chats aren’t end-to-end encrypted on platforms like WhatsApp, and Meta may use your convos to train its models—or target ads.
Protect yourself:
— Never hit “Share” unless you truly want to go public
— In the app, change settings to “Make all your prompts visible to only you”
— To stop data use for training, file an objection via Meta’s Privacy Center
— Type /reset-ai to delete AI chats across Messenger, WhatsApp, or Insta
— Bottom line: If you value privacy, think twice before chatting with Meta AI.
Meta AI’s Privacy Blunder: Private Chats Going Public
Authors
Latest from AI
Samsung’s Bespoke AI Jet Ultra reviewed
Samsung's update of a cordless classic with an AI twist gets a solid workout by our vaccy man PAT PILCHER, who finds its reputation
Enshitification: How tech companies are crapping in their own nest
We've all been the victim of it, and many of us feel entrapped by it. PAT PILCHER names and shames some of the culprits
Samsung S25 Galaxy Ultra – smarter, better, faster
PAT PILCHER is hugely impressed with Samsung's new Ultra, which refines and improves the smarts of its predecessor with the addition of some very
Dark patterns in tech: How design tricks us into doing what companies want
They're monitoring every little thing you do online and you don't know it, writes wary internet expert ANNA BUTTERFIELD. Here's her advice.
I have Insecurities about AI Therapy
Do mental health apps really care or are they merely another way to exploit the vulnerable? ANNA BUTTERFIELD explored the issue and came away
