I was driving in a car with a friend of mine back from Gdansk where I tought some cool people about iOS development and there was this story on the radio about chat bots. We were talking a bit about them and I remembered that there was this one called CleverBot which got pretty high score on Turing Test not so long time ago. I decided to talk to him/her today myself.

Now I'm convinced that we should start to make Turing Tests for people now, to get the new scale because if that one scored 42% then I'm curious how stupid those in "human" part of the test were to make CleverBot sound smart/human enough to get such a high score.

Seriously - I really, really tried to act like it was normal conversation but >90% didn't make sense. Gramatically and in any way at all. Except for places where basically any answer would be correct and places where CleverBot asked questions the conversation was similar to one you could have with a retarded, blind, eating-its-own-poo monkey with ADHD. Don't get me wrong - it's a smart monkey in a sense that it knows how to bang the keyboard with it's own head to make words sound English - that part is amazing (for a monkey). But the rest? Pretty much random selection of sentences from database would make pretty much as much sense as our conversation. I had conversations that were better and made more sense with some chat bot from a small Polish company 10 years ago than with this state of the art "computer".

I'm sure there's a great technology behind CleverBot, huge datacenters, advanced algorithms etc. but before you spread the world that computers/bots are getting smarter please check if it's not the case of just people getting dumber..