Nvidia urgently patches AI chat app after severe security flaws exposed

Nvidia’s AI chat app, ChatRTX has been revealed to have severe security vulnerabilities, but the company has been quick to issue a patch.

Nvidia’s ChatRTX app was launched in mid-February, and allowed users with RTX GPUs to build their own unique language learning model AI trained on their personal data, allowing for a more tailored experience than tools such as ChatGPT. It also ran natively on user’s PCs, so long as you had an RTX 30 or 40 series GPU.

Dexerto

However, an Nvidia security bulletin revealed that early versions of ChatRTX have “high” security issues. The first issue allows an attacker to gain access to opening files, which could lead to “local escalation of privileges, information disclosure, and data tampering.”

Another exploit, dubbed less severe by Nvidia, would allow attackers to run scripts in your browser, which could potentially lead to remote code execution, a denial of service, and information disclosure. These issues are significant and could leave existing ChatRTX users exposed to the whims of hackers, who can gain access to these aspects of your PC via UI vulnerability in the application.

However, Nvidia has been swift to issue a patch for the vulnerability, which affects all prior versions of ChatRTX. All users need to do is replace their existing version of ChatRTX with the current version available for download from their website.

With AI software rapidly evolving, many hackers have turned to AI crime to expose attack vectors in the disruptive tech sector. ChatRTX is still in beta, but users of the app should immediately update in order to protect themselves from any potential vulnerabilities.