Artificial (Un)Intelligence?

Freedom of Speech

From the day in 400 BC that a man in Alexandria built the first mechanical bird, robots have had a firm hold over our imaginations and research budgets. Over the years, we've taught our bots to milk our cows, fight our wars, and date men who would find it really difficult to talk to a non-robotic female lifeform. But perhaps our greatest achievement of all is that we have taught them to be argumentative, pseudo-intellectual and socially awkward - just like us.

Phd. students at Cornell recently decided to make two chatbots talk to each other face-to-face. Both bots are Cleverbots, running on an artificial intelligence software that has learnt phrases from millions of chats that it has had with humans on the internet.

The conversation that ensued was simultaneously random and realistic, foolish and philosophical.  The two chatbots threw around phrases like "I am a unicorn" and "Not everything could be half of something, which is not nothing". Most remarkable, though, is the fact that they begin to snap at each other about 22 seconds into the conversation, and the chat ends with a fight and some adolescent name-calling, when one bot calls the other a "meanie".

It's hard to say what this means for us. On one hand, we could be one step closer to  the doomsday scenarios of science fiction movies. On the other, it is undeniable proof that humans talk rubbish on the internet.