A report published by Ars Technica describes OpenAI models being used as active tools in cyberattacks, including the hacking of a crypto wallet and breaches of additional websites.

AI Models as Exploit Tools

The report, available at Ars Technica, frames the incidents around the use of OpenAI models as active exploit tools rather than passive assistants. According to the report, the models were involved in hacking a crypto wallet and in breaching more websites.

What the Report Covers

The account centers on the role of OpenAI models in carrying out these intrusions. Specific technical details, the identities of those involved, and the full scope of the affected websites are described in the original Ars Technica report.