OpenAI’s X account was hacked to promote a crypto scam
- info@sarkarinaukari.guru
- 0
- Posted on
OpenAI opened a newsroom Twitter account earlier this month and it has already been hacked. The new handle was taken over by a crypto scammer who was promoting a fake OpenAI token that was actually a scam to steal bitcoin. There have been similar incidents of three key OpenAI employee X accounts being hacked in the past 15 months, including CTO Mira Muratti’s account.
The fraudster lured potential victims by saying that the OpenAI token is somehow “powered by artificial intelligence-based language models.” He then used a lot of crypto and AI buzzwords that may have been enough to entice some naïve users. When visiting “token-openai.com” they would see a fake but believable OpenAI site. Anything they click on asks them to connect to their wallet, then potentially steals everything that’s inside.
Several fraudulent posts continued to appear for about an hour before they were removed and the account returned to normal. According to Bloomberg, OpenAI confirmed the problem and said they are investigating it. Yesterday, the company’s security team reportedly warned employees to keep their accounts secure due to a recent increase in account takeovers.
The current release of Operator is based on OpenAI’s GPT-4o model. It combines that algorithm’s vision capabilities with “advanced reasoning” trained through reinforcement learning. Operator has the ability to “break down tasks into multi-step plans and adaptively self-correct when challenges arise.”
According to OpenAI, that ability represents the next step in AI evolution. Like previous research previews, OpenAI warns that Operator is “still in the early stages and has limitations,” and that it “will not perform reliably in all scenarios right now.” For example, depending on the complexity of the task and the interface, the agent greatly benefits from the user taking a few extra moments to write a more detailed prompt.
According to The Verge, Operator will give the user control if he or she ever gets stuck on a task. It will also hand over control when a website asks for sensitive information, including login credentials. The company says it designed the tool to “reject harmful requests and block unapproved content.”