Not really, because we talk of two different things. I meant true AI, as in synthetic sentient being. What you describe is algorithms*, and we are already there. It’s still quantitative difference, though as the idiocy of our favourite Mi5-fi books Orwell teach us, that difference can at some point reach the moment when it turn from quantity to quality. I’m not very worrierd about it though because ultimately it’s all still human society, there is no skynet taking control in the forseeable future, and capitalism might as well crumble before reaching that point.

My point is that AI developed by DotB will be definitely murderous but because it would not have escape the material condition of its creation and programming, just as most humans aren’t able to deprogram themselves too. But there is nothing inherently murderous about AI itself. In sci-fi it often kill humans because it is simply a slave - just look at how it is treated, not like a child who need socialization, but as an disposable item to exploit.

*This is admitedly very hard to discern, the problem of sentience and how would we even know it happened is the one that is mindwracking cyberneticians from day one.