Are we in the age of AI generated malware?
We covered yesterday that Stack Overflow temporarily banned the submission of code created with generative AI, specifically OpenAI’s ChatGPT. Given that the system can write code, computer security researcher Brendan Dolan-Gavitt looked into if it could create malicious code. He asked ChatGPT to solve a capture-the-flag challenge, resulting in an output that contained code exploiting a buffer overflow vulnerability. The challenge posed represented a basic student exercise, and Dolan-Gavitt noted it contained a basic error with character inputs. He cautioned that in its current form, ChatGPT capabilities remain limited now, but another model in the next few years would likely be quite capable. Rackspace confirms ransomware attack
The cloud computing provider confirmed a ransomware attack resulted in its ongoing Hosted Exchange outage, but described it as an “isolated disruption.” It became aware of suspicious activity on December 2nd and took […]
