I'm an artificial intelligence enthusiast and I was shocked that there wasn't an AI site here. I check it out, and it seems that its beta was unsuccessful. Well, is it possible for a new start to happen?
Conceptual/algorithmic design AI questions as well as those about some practical Machine Learning concepts are on-topic at Computer Science, while anything related to statistical learning/implementation is on topic at Cross Validated. There is some overlap, so examine existing questions and tag wikis, and ask whether your question is on topic via their respective Metas if there is any lingering doubt.
Try as we might, having an independent site hasn't been in the cards. If having one is something you are interested in, rather than getting a "grass-roots" effort going on Area 51, support the existing tags on one or both of these sites with all of your might, and show that there is an unquestionable need, and audience for, separation.
There are several reasons why a beta site may have failed, and that could include timing, the specific scope of the site, activity, and quality.
This proposal didn't have enough activity during the beta. It was closed 1 year ago.
You could try proposing a variation of this site that you think will have a better chance.
My suggestion wouldn't be to go for an exact duplicate of the original proposal. That apparently didn't gather enough interest because the topic was just too narrow. Learn from the mistakes of the past. Find out why the site didn't get enough activity.
One way to approach this would be to start a new AI proposal and then get some discussion questions started to help broaden the topic without taking too much traffic from existing sites.
With that said, it's definitely okay to try to revive proposals, and it's encouraged. For example, I revived the Motorcycles proposal after it hit the 2 year mandatory close threshold.
The AI proposal failed because it had not attracted experts. This is explained in the blog post No Artificial Intelligence in Area 51.
AI (as a science) is on-topic on the Computer science site. Since the site as a whole encourages content with real science in it, there is little risk of its being plagued by the amateurish content that AI tends to attract.