Stephen Downes

Here's another take on ethical AI, this time from Miguel Guhlin. He raises the question of whether any use of AI can be ethical, given (he says) its unethical origins. He asks Perplexity for a definition of ethical AI. And he shares a longish set of prompts to prime any AI system he uses to be ethical. While I am sympathetic with the desire to have ethical AI, and to use AI ethically, I once again have the feeling that we're presuming a whole ethical analysis for all technology and all uses of technology. I mean, sure, it's easy to instruct the AI to "maintain traceability of information sources while respecting legitimate confidentiality boundaries" and maybe we should yet at the same time (a) is this an ethical principle, and (b) is it specific to AI? Why are we so much more stringent about AI ethics than, say, the ethics of coffee production?