Most ordinary people don't know about AI safety. They might have heard about 'algorithms' on the news, and they might have seen Terminator, but they don't understand what those algorithms really are, or how deep this rabbit hole actually goes. Unlike, say, climate change, there really is an awareness issue here, and it might actually matter to the outcome of things whether the public is widely aware that the risks this kind of thing poses are not just SF anymore.
Of those who do, often the ones who look into it seriously for an afternoon walk away less convinced of it being a problem than they walked into it feeling, usually justified with one of a short list of common AI risk skepticism biases (PDF). My suspicion is that, for an ordinary person, the issue seems so overwhelming, so impossible to gain traction on, that there's no sense in even trying.
I'm not convinced of this. AI safety isn't an impossible problem to manage succesfully. Its difficulty ranges between "purely collective action problem" in the worst-case scenario and "purely technical hurdle" in the best. There is power in giving people the message, "Yes it's bad - yes we can do something about it, together."
So what kinds of resources are suitable for an average, but interested, person? Like "wouldn't major in STEM" average, or even "didn't go to college for lack of ability" average.
Resources I'd be especially thankful for:
"Elevator pitch" explanations of why AI safety is a problem. Brief explanations in clear language at the 10-, 100-, and 1000-word count levels.
Common objections and their refutations to common AI risk things, since people often go from "That's not real" to "That can't happen because X" to "Maybe it won't be so bad because Y" to "Oh God wait so are we fucked?" to "No, the problem is just hard".
Specific calls to action and ways ordinary people can help, even if they can't work on technical approaches.
(more for me but I find myself jumping to here with my own friends a lot) Philosophical breakdowns of why people consider it so important that we don't end up with a paperclip maximizer in the universe. Accessible pluralist ethics, etc.
Thank you kindly, and have a wonderful day. Apologies if my English is strange in places. -Iivonen
Most ordinary people don't know about AI safety. They might have heard about 'algorithms' on the news, and they might have seen Terminator, but they don't understand what those algorithms really are, or how deep this rabbit hole actually goes. Unlike, say, climate change, there really is an awareness issue here, and it might actually matter to the outcome of things whether the public is widely aware that the risks this kind of thing poses are not just SF anymore.
Of those who do, often the ones who look into it seriously for an afternoon walk away less convinced of it being a problem than they walked into it feeling, usually justified with one of a short list of common AI risk skepticism biases (PDF). My suspicion is that, for an ordinary person, the issue seems so overwhelming, so impossible to gain traction on, that there's no sense in even trying.
I'm not convinced of this. AI safety isn't an impossible problem to manage succesfully. Its difficulty ranges between "purely collective action problem" in the worst-case scenario and "purely technical hurdle" in the best. There is power in giving people the message, "Yes it's bad - yes we can do something about it, together."
So what kinds of resources are suitable for an average, but interested, person? Like "wouldn't major in STEM" average, or even "didn't go to college for lack of ability" average.
Resources I'd be especially thankful for:
Thank you kindly, and have a wonderful day. Apologies if my English is strange in places. -Iivonen