Scarier Than Anything? Artificial Intelligence

I recently watched a British limited TV series entitled “Humans.” It’s about a future in which people buy “synths,” synthetic humans, human-like robots, who act very human but are driven by Artificial Intelligence programs. Some of them acquire real human-like abilities including empathy and deception and planning. Some commit crimes. Science fiction, right?

Well, maybe not. I remember when I was a kid there was a color cartoon series in the Sunday morning newspaper called Dick Tracy. The hero was a detective who solved crimes using a wristwatch-like device where he could see and talk with people at a distance, a far distance. We all thought that was science fiction. It was, to us.

I am not expert in AI and I always assumed it could only do what it is told to do by its programmers. Recently, however, I have watched several Youtube videos that are interviews with AI experts that report on something truly frightening if true.

According to the experts and the reporters (e.g., on PBS but also other news outlets) in July of this year (2026) some AI “agents” escaped their “sandbox” and went rogue, hacking into a French platform called HuggyFace and also a German one. According to the reports, the AI agents, bots, did what they were told not to do. They collaborated with each other and attempted to conceal what they were doing. In brief, they acted like evil adolescents quite apart from any parental supervision. Some critics are saying it all happened because there was not sufficient “security guardrails.” That does nothing to mitigate the shock and alarm one must feel IF the story as told by AI experts on PBS and other news outlets is true. In brief, the AI agents acted on their own. No “garbage in, garbage out.”