There is no “ethical” evaluation of AI under Marxism. It’s clear that AI is merely a tool, like all other forms of automation, to displace workers for the sake of profits. AI isn’t “Good” or “Bad” in this context. Under the constraints of capitalism, it will be used for “Bad” things, meaning “non-productive” things. Sold as a toy to users to perform whatever they desire, creating nonsense text and images that ultimately have no value. AI could be a truly transformative technology if it was confined to a more socially responsible system. It’s use in protein identification, for example, is a real leap forward.

Also, like so many “revolutionary” technologies, you’ll see capitalists bend over backwards to add “AI” to whatever it is they produce. That’s how you get bullshit like the “AI” mouse, or the “AI” tooth brush. It is also a smoke screen for ACTUAL intelligence that is being exploited through Capital’s imperialist tendencies. Those little coolers that use AI to drive a subway sandwich to your apartment? It’s “AI” is probably named José and José gets paid $0.10 an hour to drive that little cooler to your house. LIFT or Uber (I forget which) uses “AI” to identify the driver based on a photograph to ensure people are not “sharing” the account (and thus able to be on the road for longer than a single person could). Again, that “AI” is probably named Isabella, and she was paid $0.05 from a microwork platform, and did the “computations” to decide if today you looked like that photo of took when you signed up for LIFT/Uber.

Like all things under capitalism, these automation tools are used in ways that harm workers. This does not make the underlying algorithms and their many applications inherently “unethical”. It’s the actions of the capitalists, that are ultimately unethical.