People tend to be easily confused by things that seem vaguely similar but in reality are totally different. After all, people used to see shapes in the stars, thought trees etc. had "spirits" (many humans still think "souls" are real, continuing this simple blunder), and so forth. This is because the brain takes short-cuts in order to minimize processing time. In our ancestral environments such shortcuts were generally fine, as was our highly restricted intelligence. But this doesn't map to AI/ML in any meaningful way. The primary reason there's no true mapping is because organisms need to recognize environmental patterns (heat/cold, light/dark, etc.) that have implications for their survival and react accordingly. Organisms are DNA-reproduction machines. So survival and passing on genes are what all organisms are optimized to do. Computer programs live inside boxes. They have no sensory apparatus, no drive to survive & pass on genes, no nothing. There is thus zero comparison between animal intelligence and machine intelligence. Sure, we can train machines to do better than people at specific tasks and eventually at very wide ranges of tasks. But no AI is ever going to feel any urge to do anything in the way an animal feels urges. There's no evolutionary mechanism by means of which AI can get there. We can program pseudo-urges, but that's not evolution, that's just clever programming. And all AI, by definition, is just clever programming even though it's not 100% deterministic. Moreover, we'll be gone through self-extermination in the next few centuries, so we really don't have to worry about AI hundreds of years from now. Ordinary human stupidity will end our species, AI will just help us get there more efficiently.