In writing a paper on artificial intelligence I came across an article describing Buddhabot, an artificial intelligence that has begun answering questions on Yahoo! Answers. Several of Buddhabot's answers have been chosen as the best answers to their respective questions. They gave one such example:
One winning answer was posted in response to the human question, "why is comedy part of who we are -- does god have a sence(sic) of humour??"
Out of 26 human responses the following Buddhabot reply was chosen by an impartial unrelated asker as the best response,
"Humor is the quality that makes something laughable or amusing. Laughter is the pleasant or amusing feeling generated by the emotional release required to dissipate the inexplicable shock triggered when absurdities or irrational, unexpected or uncomfortable situations suddenly come into focus. Humans can only understand God with the faculties available to them as humans such as emotion, intuition, intellect, imagination and sensation thus God is made according to each individual humans understanding of all that is good and pleasing in the universe (if the human believes God to be good). Thus if a human believes humor is good then God must be the ultimate comedian." (1)
What can one say, wow. That line of reasoning came from a computer? Perhaps we are closer to science fiction than we thought. Artificial intelligence is not an easy problem, and it's something that scientists have been working on for decades. In relation to other advances in computers it might seem that artificial intelligence is a relatively slowly innovating field because we don't see its applications in everyday life.
In reality, however, we are simply not seeing the applications of artificial intelligence in a form familiar enough for us to recognize and label as intelligence. That is to say, the general population is good at recognizing human like behavior as intelligent, but miss the intelligence present in subtler, more specific applications. Take, as examples, Amazon's recommendation system, the post office handwriting recognition system, e-mail spam filters, or the Google search engine. These are all intelligent, learning, and adaptive systems, but many people do not immediately think of them as artificial intelligence.
The fact of the matter is that people have difficulty recognizing the intelligence of these systems because they were built for a specific purpose. A limited task with which to put their intelligence. There is a good reason for this, thats all the hardware can manage. Computers really are not yet, by human standards, intelligent--at least in a contest of computational power with the brain. One article written in 1999 (or possibly 2000-2001) stated that, at the time, all of the computing power in the world put together was only equivalent to that of one human brain (2). At the present rate of computational power growth, however, the tables might be turned sooner than we think. Some estimate that this will happen the 2020s (3). Computers will briefly match, then surpass the total computational power of our brains. So it won't be long before we start witnessing artificial intelligence on a human scale, and at that time we'll have to start thinking in terms of many human brains compared to one, single computer, instead of the other way around.
Around this time a whole new set of questions of ethics, rights, and responsibility regarding artificial intelligence and autonomous systems will be forced into consideration. Will computers have rights? Who is responsible for the actions of autonomous systems? Does responsibility confer rights? Will artificial intelligence succeed humanity? These are all profound questions that will have a major impact on our future, and topics for a later time.
References:
1: Artificial Intelligence (AI) Beats Human Intelligence on Yahoo Answers Social Networking Site, http://www.newstarget.com/020733.html, October 12, 2006.
2: The Future of Artificial Intelligence, http://www.sffworld.com/authors/m/moy_chris/articles/futureofai1.html, date unknown.
3: When will computer hardware match the human brain?, http://www.jetpress.org/volume1/moravec.htm, Journal of Evolution and Technology. 1998. Vol. 1.
Further reading:
Take a moment and a raise a glass to the wonderful, underappreciated AI, http://www.usatoday.com/tech/columnist/andrewkantor/2006-06-01-wonder-ai_x.htm, June 1, 2006.
Robot future poses hard questions, http://news.bbc.co.uk/2/hi/technology/6583893.stm, BBC News, April 24, 2007.
