U.S. Markets close in 6 hrs 2 mins

IBM unveils a computer that can argue

Rick Newman
The Exchange
FILE - In this Jan. 13, 2011, file photo, "Jeopardy!" champions Ken Jennings, left, and Brad Rutter, right, flank a prop representing Watson during a practice round of the "Jeopardy!" quiz show in Yorktown Heights, N.Y. Watson, the question-answering supercomputer is going to college. IBM is announcing Wednesday that it will provide a Watson system to Rensselaer Polytechnic Institute, the first time a version of the computer is being sent to a university. (AP Photo/Seth Wenig, File)

I think, therefore I argue.

This might be a more apt summary of humankind’s intellectual evolution than Descartes’s famous dictum, and computers are now treading the same path. IBM (IBM) has been a leader in developing artificial intelligence systems such as the Watson supercomputer, which won $1 million on Jeopardy in 2011, beating out top human champs. And now Big Blue has taken Watson technology one step further with a system that can form logical arguments for or against a complex issue--rather than just answering questions--once it absorbs relevant information.

At the Milken Institute’s annual conference in Beverly Hills, John Kelly III, IBM’s director of research, unveiled an artificial intelligence project called the Debater that has new capabilities to think—and argue-- like a human. In a demonstration, Kelly asked the Debater to provide arguments for and against this statement: The sale of violent videogames to minors should be banned. The contraption scanned 4 million Wikipedia articles and narrowed them down to a handful that seemed most relevant. Then, in a synthesized voice, it spit out answers.

Among the reasons to support a ban: “Exposure to violent videogames can cause adolescents to commit acts of aggression in real life,” the Debater explained. ”On the other hand,” it went on, “the sale of violent videogames to minors has not been causally linked with aggressive tendencies.” Doing the research and rattling off several statements, pro and con, took about 45 seconds.

The technology breakthrough here is the ability to understand spoken and written language in context, then applying computing power to a question that arises from that language. The human brain is far better at contextual learning than computers, because we can almost instantly ascertain whether a word like “fair” pertains to a square deal or a carnival with a Ferris wheel. A few minutes of speech or a few paragraphs of writing might contain dozens of such associations, leaving a typical computer stupefied, no matter what its processing speed or RAM.

IBM isn’t just showing off a smart-alecky machine on a whim. The Debater, which is still a research project not yet commercially available, could have plenty of important real-life uses once it comes to market. Medical researchers could use its services to get a big head start when they’re trying to find cures and treatments for diseases. Its ability to summarize terabytes of data in ways that are logical to humans could allow it to summarize all the research that’s been done on a given disease, while also identifying new and unexplored avenues of inquiry that could yield promising discoveries. The same goes for legal research, military intelligence and many types of science and technology.

Kelly envisions the Debater having a seat at the table — either in the flesh, so to speak, or through the cloud — when people in disparate fields develop projects to solve complex problems. It can listen to conversations and contribute meaningful information through real-time data crunching. The Debater also learns as it listens, adding information gleaned from its human interlocutors to its gigantic database. “It’s not man versus machine,” Kelly insists. “It’s man and machine reasoning together.”

On an individual level, the Debater could help develop a personalized course of treatment for cancer patients by processing all known factors associated with the individual case, and comparing that with all the data available on treatment outcomes for different types of malignancies. That might allow doctors to prescribe a precise cocktail of drugs or other treatments with the highest likelihood of success, avoiding a trial-and-error approach that often burns valuable time. If there's a risk, it might be that practitioners in the future come to rely too much on machines, taking human judgment out of the process.

The next step will be adding visual capability to such a machine, so it can see what’s going on around it and add images to its databases and reasoning power. Possible uses include military combat, space exploration and other realms where a sentient machine might be preferable to a human.

Here’s a question for the Debater: What will happen as artificial intelligence replaces human reasoning? Pro: There’s a long history of new technology giving humans fresh tools to become even more creative and productive. Con: When the robots have taken over, there may not be any jobs left for humans. Maybe the next IBM supercomputer will be able to feel bad about that.

Rick Newman’s latest book is Rebounders: How Winners Pivot From Setback To Success. Follow him on Twitter: @rickjnewman.