So I was in a meeting at work the other day and one of the topic was "Tags". Tags are becoming quite popular on content heavy websites these days and they are quite useful to boot. The thing about tags though is finding the right way to control them. As far as I can see, there are two ways to do this:
The first - and easiest - is to have your users create them as needed. A draw back is that you may get duplicates, similar tags, spelling mistakes etc. which will cause your tags as a whole to be kind of useless to the end user. For example, a user accidentally tagged their post with "blckCat" and forgot the "a". Now, anyone looking for blackCate won't be able to see that post.
The second is to generate your own tags and then have your user choose among them. This solves the previous problem, however creating each possible tag for your site could potentially take a really long time - in some cases may even be impossible.
Then I began thinking... what if we were to have a sort-of "Ai for tags"? An entity that would work almost like as a validator for tags as they are created. It would be responsible for knowing all words in the dictionary obviously, but it would also learn new words as tags are created. And, it should be smart enough to know when a tag is being created that is similar to an existing one or if there is a misspelling of a currently existing tag.
... just a thought