Heidy Khlaaf, a chief AI scientist at the AI Now Institute, said she remains skeptical about the seriousness of the hack because of a lack of detailed information from both companies about their security posture. She said the situation was rooted in giving the AI “a task with poor parameters,” versus the AI acting without human control.
“These types of environments are actually notoriously very insecure,” Khlaaf said of the experimental environment containing OpenAI’s model. “They’re easy to bypass, with dozens of incidents like this being already reported in the past. So to me, that part of the sort of puzzle
wasn’t new either, right?”
“And so this is why we need much more information to really understand the severity of it,” she
added.
Read the article here.
Research Areas