Read Complex Worlds from Simpler Nervous Systems[1] if you want more information on how very complex behavior can emerge from jumping spiders and bees. After reading that book, it seemed to me that AI researchers should focus much more time on duplicating the feats of these simpler animals (in similarly computationally limited contexts) rather than focusing on duplicating extremely high level human faculties like reasoning and even playing Go. I mean, imagine a robot with the intelligence of a parrot or the smartest birds in the corvidae family[2]. I think people would probably be afraid of autonomous robots of this level of intelligence.

[1] https://mitpress.mit.edu/books/complex-worlds-simpler-nervou...

[2] https://en.wikipedia.org/wiki/New_Caledonian_crow