A computer science professor at UW-Madison says certain AI tools pose a cybersecurity threat when in the wrong hands, and urges skepticism about potential defenses.
In a presentation held as part of Madison’s Forward Festival, Prof. Somesh Jha yesterday discussed how experts in the field are looking to preempt the use of these tools by hackers and other adversaries. Some AI programs are likely already being used in this way, he said, pointing to large language models like ChatGPT that can create sophisticated written content.
“But the problem is that we don’t have a good way of understanding where is that threat landscape right now,” he said. “If you talk to people who are in this threat intelligence community, they’ll tell you that, yeah you’re seeing already evidence that they have started using it, quite heavily.”
;
