I.B.M.’s Watson is a remarkable achievement, beating the two best Jeopardy champions. Its success shows that significant progress can be made in processing natural language by statistical analysis of huge amounts of text. Its occasional startling errors show that statistical analysis of text is insufficient to really understand the meaning of language. For that, it is necessary to analyze the relationship of language with sensory and motor experience. I am confident that some future Watson with eyes, ears, limbs and fingers, and a brain able to statistically analyze the data from its body in relation to text, will truly understand language. It will understand the word “chair”, not merely from occurrences of “chair” in text, but from its experience seeing a wide variety of chairs being used, lifting and carrying chairs, and possibly building and dismantling chairs. The combination of such physical experiences for thousands of words will give it real language understanding.
Despite the current Watson’s lack of real understanding, it was able to beat the best humans at a language game, because its language processing ability is connected to a memory of names, dates and facts that no human can match. When a future Watson can understand language, it will be connected to an even larger database, and to the ability to run state-of-the-art simulation models of weather, chemistry, biology, economics, and many other domains. Just as the current Watson can beat any human at Jeopardy, the future Watson will be a potential champion at the real-world language game called politics. It will skillfully frame issues, debate, call and poll voters, write position papers, and issue press releases.
The Jeopardy-champion Watson runs on an expensive computer, and I.B.M. spent much money on developing the architecture and software. A future Watson, with the computing capacity to statistically analyze sensory and motor experience and its relation to language, will require a very expensive computer, at least initially, and a huge effort by well-paid scientists and engineers. This expense will be borne by some group with a serious motive for having such a machine. Perhaps they will be a corporation like Google, Facebook or Twitter with huge amounts of user searches, posts or messages to analyze and understand. Or perhaps they will be the U.S. Government, sifting through the world’s email and phone calls looking for terrorists. In any case, it is likely that the future Watson will know a lot about each of us. This could be quite useful if the future Watson is employed for politics.
Increasingly, smart machines are being used by some businesses to reduce their human workforces, and by other businesses for complex investment strategies. The resulting unemployment and financial instability are hot political topics, which is a motive for the businesses employing smart machines to use those machines to help push their political agendas. Google’s motto is ‘Don’t be evil’, and I believe they mean it. And I think the people running the U.S. Government generally have good intentions, even when I strongly disagree with them. But people get used to wealth and power, and come to see them as their right. And people in power are often willing to bend the rules in response to perceived threats.
Public fears of AI seem to be focused on the threat of machines killing or enslaving humanity, as in movies like The Terminator and The Matrix. In response, AI experts have been working to calm their fears. In its August 2009 Interim Report, the AAAI (Association for the Advancement of Artificial Intelligence) Presidential Panel on Long-Term AI Futures wrote “The group suggested outreach and communication to people and organizations about the low likelihood of the radical outcomes.” Ray Kurzweil is the leading public voice about AI. In his 2005 book, The Singularity Is Near, trying to calm fears of AI, he wrote that it will be “intimately embedded in our bodies and brains”, and hence “it will reflect our values because it will be us.” Jaron Lanier, in an August 2010 New York Times op-ed about AI, wrote “Technology is essentially a form of service. We work to make the world better.” I believe we need to focus on the threat of AI as a tool for competition among humans, and in particular, the threat of AI employed for politics. We need to inform the public that this is the real threat, rather than the dramatic “AI versus humanity” threat seen in fiction and movies.
Humanity+ is about the future ability of humans to enhance their bodies and brains for improved health, indefinite longevity, greater physical skills and increased intelligence. This will be an unprecedented benefit for humanity. But it will alter certain realities that underlie human society. In particular, most humans have roughly the same intelligence. We all speak the same languages. If one person becomes an expert in a field with specialized language, another person can learn it. But humans whose brains are enhanced with the technology of the future Watson will have language skills that enable them to speak languages too complex for natural humans to ever learn. An enhanced human brain will speak a language that combines and exceeds all current specialties, is informed by instant access to all the facts of Wikipedia and all scientific simulation models, and incorporates new insights equal to those of Euclid, Newton, Darwin and Einstein in every expressed thought. This is a glorious vision that we all aspire to. But those unable to understand the language of the most intelligent will be unable to participate in serious discussions of public policy. When someone’s intelligence depends on how large a brain they can afford, the current economic inequality among humans will translate into extreme biological, intellectual and political inequality.
Artificial intelligence and human enhancement are goals that we must attain to make a better world. I.B.M.’s Watson is a champion at the language game called Jeopardy. Politics is a more serious language game, and we must carefully deal with the possibility that some future Watson will be a champion at politics, in order to protect the benefits offered by enhancement.