Proposal: Artificial Intelligence

Personally I would prefer mainly technical. It seems that a few of the questions involve semi philosophical, and somewhat subjective questions.

  • Technical, yes. Philosophical, probably. And ethical, which I think would be as interesting. – Yisela Apr 30 '16 at 21:01

It should be both.

On a new site with technical questions only, most of the questions would be on-topic on the Data Science portal: https://datascience.stackexchange.com/

And I think with mostly theoretical questions the site would have not enough questions/visitors.


I also think it should be both. Here are other reasons to build on @Iter Ator's answer.

  • AI is not only Data Science, and the idea is to offer a space to discuss other dimensions (please refer to this answer for examples).
  • Theological may be a strong word, but yes, AI is also about the responsibility of becoming a "creator", when we talk about the so-called "strong AI". OpenAI and related initiatives like the FLI (and its famous open letter) are examples in this category.
  • Theoretical is important, for example considering that AI does not need to mimic human intelligence. There are other ways to observe "intelligent" phenomena, e.g. swarm intelligence and emergence. These are "models", and quite different from creating a "Clipper" assistant...
  • I was reading this article today about Snapchat's ethical responsibility for creating the "speed filter" that causes so many car accidents. Although I can clearly see the need for a discussion about corporate responsibility in this case, I wouldn't call Snapchat's issue a theological one. I'm trying to get my head around the concept - the questions can get theological, but the development is not. Me thinks... – Yisela Apr 29 '16 at 13:03

I think it would have to be only theological, and not technical. If it was technical, then a lot of questions would end up being duplicates of existing websites that already handle technical questions (such as "data science").

The philosophy about AI is remarkably separate from actual AI research/development, and it still remains that way because "artificial intelligence" is itself a philosophical concept, dealing with vague concepts such as, you know, the term 'intelligence'. The type of people who are interested in building AI are not necessarily going to be the same people who want to learn about AI and make predictions about the future. There's already lots of resources for the former. There's very little (informed) resources about the latter.

I don't worry about people getting bored asking about AI philosophy and leaving the website. AI is not going anywhere, and neither are disputes over the definition of 'artificial intelligence'. There will always be demand for somebody to ask and answer questions.

  • 1
    I don't think that would gather enough people to work successfully. I for one would not use it at all, as amI mainly interested in actually getting stuff done. Overlap between communities is ok. – Raven Mar 28 '16 at 3:23
  • History of AI as a science shows quite a few techniques that do not seem to fit on Data Science. For example, rule-based systems (should I say "expert systems"), case-based reasoning, automated negotiation, etc. The overlap will exist for DS and Statistics-related topics such as many ML techniques, but that should be fine (to some extent). Like it works between Seasoned Advice and some other food related SE sites (e.g. Coffee). How does this sound? – Eric Platon Mar 30 '16 at 1:00

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .