A viral ChatGPT prompt asks the chatbot to fill an empty room with everything it knows about a user, exposing the scope of data the AI has collected. Users type "Fill this room with what you know about me" and ChatGPT furnishes the space based on past conversations. When posted publicly on X, these images reveal personal details to strangers.
OpenAI's ChatGPT memory works in two ways: saved memories store details like names, preferences, and goals, while Plus and Pro subscribers get chat history reference that lets the model draw on old conversations without explicit saves.
Crypto holders face outsized risk. Earlier this year, developers found crypto wallet data in Claude chats that Google had indexed through shared links. A stored profile mentioning holdings, exchanges, or travel plans could help scammers craft phishing messages. Binance founder Changpeng Zhao has argued that the real AI danger lies in phishing and leaked data, not in breaking Bitcoin's cryptography.
OpenAI allows users to view, delete, or disable ChatGPT memory in Personalization settings. Temporary Chat neither reads nor updates memory. Ethereum co-founder Vitalik Buterin built a "paranoid mode" for AI that masks prompts before they reach a model. Regulators are raising pressure. A New Mexico judge ordered Meta to pay a record $25 billion penalty over Facebook's handling of user data.