ChatGPT Malware: Hackers use sharing feature to spread it

Attackers are using ChatGPT's content-sharing feature to spread malware. They create fake OpenAI outage pages, tricking users into downloading malicious software disguised as a desktop application. This campaign, called "LLMShare," is still being detected by security researchers. What's interesting is how convincing these fake pages are. They're hosted on ChatGPT's own domain and abuse its code rendering feature to display a service disruption message. This message prompts users to download a desktop app to continue. The download button leads to a clone of ChatGPT's official download page, complete with OpenAI branding. This site uses cloaking, showing a fake download page to users but a benign one to security scanners, making detection harder. Similar attacks have also been seen on Claude. This method relies on user trust, as the fake outage page appears on a trusted domain. The bottom line: this type of attack can lead to stolen credentials or remote access, posing a significant risk to organizations.