I talked with an AI today.
Salon had an article on a chatbot that was the most successful in fooling judges in the annual Loebner prize. So naturally I had to check it out.
It’s name is Elbot. It’s by turns cheeky, philosophical, and obscure. Really, really obscure. James Joyce has nothing on Elbot. But what was interesting was the way my own side of the conversation evolved. I was friendly, chatty, open — in fact, I was a better conversationalist than I usually am with people I’ve just met. I wonder what that says about me, or even if the Turing test measures the humanity of the human side of the interaction. Even though Elbot was mechanical and canned, I was more relaxed and open. (If you’ve ever met me in real life you know that I am in fact dour, uncommunicative, and surly.)
So maybe AIs shouldn’t become more human. Maybe they should stay the way they are — awkward, a little off, kind of weird. And us humans could become more — human. Maybe that’s what’s supposed to happen. Machines will stay machines and humans will find humanity.
Chatting with Elbot reminded me of another AI conversation, called Gimpbomb Enters The Room, by Matthew Bey. Elbot is where we are — does that mean Gimpbomb is where we are going?
1 Comment
amichail · October 14, 2008 at 11:48 am
Check out this Web 2.0 approach to chatbots: http://chatbotgame.com.
Just as Deep Blue brute-forced it in chess with speed, the idea behind the Chatbot Game is to brute-force it with a huge number of user-submitted Google-like chat rules.