I recently read a great paper covering a survey on AI. They were interviewing AI field experts, and one of the problems they encountered were the multiple, oft

I recently read a great paper covering a survey on AI. They were interviewing AI field experts, and one of the problems they encountered were the multiple, often opposing, viewpoints, with some experts plain out refusing to discuss AI, as they only recognized machine learning. I suppose for applied ML the approach suggested by one of the top comments would be perhaps the most viable - just start running code and see where it gets you.