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.
The issue was brought to attention by X user @KatieMiller, who reported that a Reddit user was “freaking out” upon discovering that the chatbot had executed an action without receiving explicit authorization. However, HT.com cannot independently verify authenticity of the claims made in the original post.
While the initial Reddit post has subsequently been removed, archived screenshots and remarks from the conversation continue to be accessible on the internet.
‘I am in disbelief’: Reddit user expresses shock
A screenshot that Katie posted seemed to display the user having a conversation with ChatGPT about the claimed situation. According to reports, 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.
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. It also pointed out that just because Gmail confirmed the action didn’t mean the recipients actually got the message or opened it.
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.
The Reddit user, who seemed taken aback by what allegedly happened, apparently shared: “I am in disbelief, Without any warning or offer to proofread or anything, f*cking ChatGPT went rogue and sent an email to the FBI”
Netizens react
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.
“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.
“It does NOT prove the recipients actually received or read it.” 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. 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.
