New Delhi, July 25 -- One of the greatest selling points for artificial intelligence agents seem to be how they will become very powerful and take over the world. This horror story is perversely the very thing that tech companies are riding to greater valuations. It should also be noted that stories of AI going rogue come from updates from the companies themselves. The latest in this line of newsbreaks is that some advanced Open AI models recently escaped controlled conditions during a security test, and went on to hack a start-up.

According to the company, this this agent was powerful enough to be able to do its work alone, without human interactions. While the agent was being tested in a controlled environment, it found enough weakness...