AI based comment authentication

PH went ahead and started his own webpage, and its pretty cool. The one thing he debated on was whether to have comments or not, and after careful consideration, he decided not to.

Well, I have had comments turn off for over a year, as well as trackbacks, due to the huge amount of spamming that was going on. I do really miss comments, but the spammers ruined it for everyone.

I got to thinking on this, and even with Captchas and questions spammers sooner or later find a way to break in and cause havoc, and also, such programs tend to discourage comments, due to the extra hoop jumping.

Then I had a thought, why not analyze the comments content for relevence to the articles content. If there is a reasonable level of correlation, then accept the comment. If not, go into massive hoop jumping mode, where rather than a simple question, and AI question based upon the articles content.

By doing so, comments are encouraged, and spammers are for the most part, either relegated to reading and comprehending an article, or they will walk away.

Of course, the difficulty is doing so. The AI classes I took 20 years ago were pretty cool, even though we had to write our code for a mainframe computer. Now, with all the computing power available, it would seem an almost no brainer. I wonder why no one has done so.