What happens when two chatbots chat? Researchers from
Cornell Creative Machines Lab headed by Professor Hod Lipson at Cornell University have recently submitted this entertaining idea to this year’s Loebner Prize Competition in Artificial Intelligence. Objective of this competition is to fool 2 or more judges into thinking that a chatbot is a human being.
Cornell’s Creative Machines Lab experiment of AI vs. AI is based on three components: a chatbot, a text-to-speech synthesizer, and an avatar renderer. Initially, the researchers used Eliza, which tended to produce quite boring conversations, therefore they switched to Cleverbot, a smarter and constantly learning chatbot. The text-to-speech synthesizer takes the text generated by the chatbot and creates a spoken, audio version. To synthesize an animated character whose gestures and lips are synched to the sound stream an avatar renderer was used. By combining these 3 components in Python, a single machine was produced that can converse with a user. The output of one machine was plugged into the input of a second, and the output of the second back into the first, producing endless comic robotic entertainment (or genuine artificial intelligence?...).
As you can see in this video below, the conversation between the two chatbots becomes quite funny, ranging from saying hi to discussions about God and its existence.