ChatGPT Privacy Settings Most Users Don't Know Exist (Turn These Off Now)
The short version: Mading Thok's walkthrough of hidden ChatGPT privacy toggles is a reminder that opt-out defaults, not opt-in consent, define most modern surveillance — digital and physical.
Our Take
Credit to Mading Thok for doing the unglamorous work of digging through OpenAI's settings menu so the rest of us don't have to. The video walks through four controls — model training, Memory, Temporary Chat, and full data deletion — that most ChatGPT users never touch because they don't know they exist. That's the pattern worth noticing: the default setting is almost never the privacy-protective one, and the burden is always on you to find the switch.
That same logic is exactly how automated license plate reader networks operate. Flock Safety cameras get installed on public roads, log every passing vehicle, and build location histories on residents who never opted into anything and often don't know the hardware exists until someone points it out. There's no toggle buried in a settings menu for a chatbot; there's a camera bolted to a pole, feeding a database that police departments and sometimes private entities can query. The consent conversation ChatGPT users are finally having about their prompts is a conversation most communities haven't even started having about their license plates.
If you're the type to go digging through an app's privacy settings, don't stop at software. Check our camera map to see what's logging your plate on your own streets, and use our take-action resources to push back at the local level — because unlike a chat history, you can't just hit delete on a year of ALPR location data once it's been collected.
This is DeFlock The USA’s original commentary. The video above is the work of Mading Thok, published on YouTube — full credit to the creator.