Yes, machines are not as complex as animals. But already they have been able to produce more intelligent-looking responses than any animal except a human.

My mind has already been trained on images of dogs, uses of the word "dog" and so on, which is why I am able to understand "don't think of a dog" and able to imagine a dog. An LLM is different in that it can't imagine anything, but it's also similar in that it doesn't look up images or references when you ask it a question - it just runs the statement through its neural network which has already been trained on the images and words, just as we run the statement through our minds which have been trained on images and words.

There really is no possibility in which AI develops near-human aptitude for science yet we retain our freedom without a near extinction event. Either the AI is widely available and any psycho can ask one to make a weapon of mass destruction or the AI is controlled by a few who will have unchecked power over everyone else in the world. This post breaks it down a bit further: https://scored.co/c/StopTech/p/1ARJcvK5JL/why-technology-must-be-stopped/c