I think you're falling prey to having limited knowledge about a field and then expanding on that knowledge thinking that what you think must be new because you didn't know it before.

Probably the best course of action if you want to advance the field of AI is to first get current with the the field, then start thinking up new stuff, preferably in concert with other people in the field to keep you on the right track. It's rare that an outsider with limited knowledge of a field can suddenly make some fairly obvious contribution that those in the field had overlooked.