I have recently been interested in robotics, particularly in for-profit startups. I think they can help increase food production and help reduce improve healthcare. Would this fall under AI for social good? How impactful will robotics be to society? How large is the counterfactual?

Thanks for posting this question. I'll take a stab at it.

I think they can help increase food production and help reduce improve healthcare. Would this fall under AI for social good?

Yes, any use of AI that aims to improve society counts as AI for social good.

How impactful will robotics be to society? How large is the counterfactual?

It's hard to know because this question is so broad - one could spend their entire life investigating it. It would be helpful to narrow down your question into something more specific, like "the projected future impact of agricultural applications of robotics on human well-being". Then you could divide agricultural applications of robotics into different use cases and predict how each use case would affect the outcome you're interested in (in this case, well-being) and how much. Also, the counterfactual impact of an individual new technology (such as robotics) is hard to isolate, because technologies are very often used together in different combinations (such as robotics + computer vision + hardware accelerators for AI + ...), and because technologies interact with other technologies and other social systems in complex ways.

Once you've narrowed down to a more specific question, you've got something you could do an entire year-long research project on. We wouldn't be able to answer your research question on the EA Forum overnight, although we may be able to point you to relevant resources if you ask a specific question.

Finally, a shameless plug: if you're interested in using AI to improve the world as effectively as possible, I invite you to join my Slack.

I hope this helps!