A lot of bad thinking about AI flows from inappropriately anthropomorphizing AI systems. Examples:

(1) The Copyright Office's stance that you can't copyright AI-generated work flows from the mistaken idea that an AI is in some sense a person and can therefore be the author of a work. Actually Midjourney is a tool like photoshop and the human user is the author in both cases.

@binarybits Moving away from statistical systems, which steal content and aren’t actually intelligent in any way, will help a great deal here.

@binarybits Indeed. We cover multiple facets in our paper that featured in EMNLP: arxiv.org/pdf/2305.14784…

@binarybits You're right that anthropomorphizing AI is a problem. But some of your arguments are backward. LLMs are not tools like Photoshop; they're tools like copy machines. "Prompt engineering" is a joke job. The significant human inputs come from the training material.

@binarybits This is true, though also a lot of bad thinking is the opposite – inappropriately refraining from anthropomorphizing. For instance, I think one major error of the "stochastic parrots" crowd is refusing to consider hypotheses that look like anthropomorphizing.

@binarybits ChatGPT is "thinking".

@binarybits 1. You’re right for now. 2. I really want my own agent 3. In the beginning of LLMs I didn’t know the limitations in language. 4. No, it’s evolutionary force that will stem from humans, that will create that risk. All these debates, is the e/acc way forward to a bright future.

@binarybits Totally agree. The “bewitchment of language”

@binarybits It's not even anthropomorphizing software. It's maybe biomorphizing? Just thinking that software is alive, with its own (mostly) independent goals and choices for how to achieve them. Software isn't even as functional as an amoeba. It's a great tool, but that's all.

@binarybits A lot of bad thinking about AGI/ASI flows from refusing to apply "anthropomorphic" common sense to conscious machines, systems that are likely to be quite anthropomorphic in design and in the existential predicament that they will find themselves to be in.

@binarybits Tell me about it

@binarybits On point

@binarybits I agree with you that the anthropomorphisation of AI is done by high school level philosophers with a bad grasp on psychology and history

@binarybits Whether anthropomorphizing is inappropriate depends on whether the machines are actually like us in the relevant respects. In some cases - nursemaid, companion, geisha - the machines could come to be be quite like us. Then, anthropomorphizing won't be particularly inappropriate.