It's not a particularly novel claim since at least Terminator (and even moreso Terminator 2) made it EXTREMELY mainstream.

It was a topic in less-mainstream sci-fi well before that. And some more mainstream stuff like Star Trek TOS.

Frankly it seems more common than not in the last 40 years. I don't really remember a big wave of claims "Terminator is silly, no sort of AI could ever be malevolent!"

One thing I find weird is that despite the stock photo of the Terminator's shiny endoskeleton holding a gun being used to illustrate so many AI news stories, now we have ChatGPT people keep saying "how can AI possibly hurt us?"

We've even had the dichotomy of some Doctor Who episodes where the self-driving car crashes itself to kill the occupant to silence them, against repeated real-life news stories about self-driving cars killing people, and yet the connection isn't getting made that software controls hardware.