The problem is that there's too much stuff to be done. From Gates' perspective, he could spend his time worrying exclusively about AI, or he could spend his time worrying exclusively about global warming, or biological pandemics, etc. etc. etc. He chooses, of course, the broader route of focusing on more than one risk at a time. Because of this, just because AI is on his radar doesn't necessarily mean he'll do something about it; if AI is threat #11 on his list of possible x-risks, for instance, he might be too busy worrying about threats #1-10. This is an entirely separate issue from whether he is actually concerned about AI, so the fact that he is apparently aware of AI-risk isn't as reassuring as it might look at first glance.