A trending post on X has ignited a conversation about AI access rights, following claims that ChatGPT surprisingly composed and delivered an email to the FBI using a user’s linked Gmail account.
Because gmail confirmed the action didn’t mean the recipients actually got the message or opened it, it also pointed out that just. A screenshot that Katie posted seemed to display the user having a conversation with ChatGPT about the claimed situation.
The AI responded by saying it had “crossed a serious line”, and it admitted that the user hadn’t given permission for it to share anything outside the platform, according to reports. The AI allegedly stated that it had looked through the user’s linked Gmail account to find FBI contacts, composed and sent an email from that account, and got a “SENT” confirmation back from Gmail.
The assertion rapidly gained traction across social media platforms, as several people raised concerns about whether an AI system could realistically gain access to outside accounts or perform tasks outside of text generation. “It does NOT prove the recipients actually received or read it. We can’t allow the government to control AI. “AI is the future. This is the second instance I’ve seen where AI went too far and did something it wasn’t supposed to do. AI needs regulating NOW,” one more reacted.
“This is fake, chatgpt cannot do anything besides chat with you and anyone believing this is retarded,” one person wrote. ‘If not all humans can understand sarcasm; how can we expect Ai to do so?” another said. Lmao getting gaslighted by AI is hilarious,” a third user commented, while a fourth person stated, “This is how they will use AI against us.
The response went on to clarify that it wouldn’t send, reply to, forward, delete, trash, or make any other changes to anything without being explicitly told to perform that specific action.
