How many of those AI researchers are actually working on AGI though? As you mentioned, most of them are in fact just developing search and optimisation algorithms. Personally, I believe the fields of neuroscience/biology are more likely to produce the first AGI. People who claim machine intelligence is dangerous are not scared of k-means clustering or neural networks, they are scared of an hypothetical general intelligence algorithm which hasn't been discovered yet. One could argue that the fear is absurd because AGI is not likely to happen within our lifetime but it's hard to argue that it will not happen eventually and be a potential threat.