AI models are inching closer to hacking on their own

The new report indicates this fear could be a reality sooner than anticipated.

Read at Axios