A vulnerability was found on ChatGPT.

A vulnerability was found on ChatGPT when summarizing the website by AI. Phishing scam can be used.

According to a report by The Hacker News website, a vulnerability has been detected in the web summary of the ChatGPT. ChatGPT can be used in a new type of deception technique called ChatGPhish. ChatGPT believes in the Markdown section. The URL and the image on the website of the hackers will be used in a new type of Phishing campaign. Using the ChatGPT technique, summarize the website by pulling (Fetch), the link section, and the markdown image on the website of these hackers in a screen for users to interact with. Or a UI (User Interface) that looks at ChatGPT trusted pages.

It is easy to summarize this: Hackers simply write code in the form of a display on their website and then command ChatGPT to summarize the page. ChatGPT will then extract the code, links and images on the page that the hacker has made to create a credible screen. This can range from creating a fake notification page and making a QR code to deceive the victim with great ease. In this way, hackers can insert the above commands on the page of the hacked website and then after the victim is trapped in the page. The command to summarize the page. The Prompt or hidden command on the page will be reconstructed on the screen of ChatGPT. The victim actually belongs to the website, but when pressed back and led to malware infection, it can be called a launch attack or Prompt Injection.

But users may be relieved that the method is only a test of a research team of cybersecurity experts, whether it works, or whether the AI developer will complete the correction first, perhaps to follow up.

# Trending # Lemon 8 Howtoo # lemon 8 diary # chatgpt # freedomhack