On May 11, 1997, an IBM computer called Deep Blue defeated the reigning world chess champion, Garry Kasparov, capturing the attention and imagination of the world. The six-game match lasted several days and ended with two wins for IBM, one for the champion and three draws.
The match was cast as the ultimate example of man versus machine. Yet, Kasparov himself is now a self-described proponent of artificial intelligence (AI), recently calling it a boon to humankind, “capable of providing us with endless opportunities to extend our capabilities and improve our lives.”
This view mirrors IBM’s. We too believe the promise of AI lies in its capacity to augment human abilities and intellect.
Over the last 20 years, IBM has worked to advance the field of AI. Deep Blue used algorithms to explore up to 200 million possible chess positions per second, then chose the move with the highest likelihood of success. While Deep Blue did use machine learning approaches, it relied primarily on a programmed understanding of the game of chess – 64 squares, 32 pieces and well-defined moves and goals.
IBM Deep Blue team members wheel the system through the streets of New York City to Rockefeller Center at 4:00AM, May 11, 1997.
Fourteen years after Deep Blue’s win, IBM applied AI to the more dynamic real-world challenge of Jeopardy! This represented a huge step forward from playing chess, given the quiz show could cover questions on just about anything. IBM Watson incorporated facets of AI, machine learning, deep question answering and natural language processing to play and, ultimately, bested the game’s greatest human champions.
Much was learned in pitting AI systems against human opponents. Most significant is the realization that AI can best be used to fill the gaps in human ability, and vice versa. AI technology has evolved significantly, but AI systems are not nearly as good as humans at common sense reasoning or thinking creatively and strategically. Such gaps will likely remain for decades to come.
The Deep Blue Team 1997: Chung-Jen Tan, Murray Campbell, Joe Hoane, Feng-Hsiung Hsu, and Jerry Brody.
Therefore, we must factor the abilities and limitations of both machines and humans to create combined systems that enable outcomes better than what either could achieve alone. The promise of AI lies in the combination of man and machine.
IBM Watson is on track to be used in some form by a billion people by the end of 2017. In the next five years, many important decisions, be they business or personal, will be made with the assistance of a cognitive system like Watson.
IBM’s goal is to accelerate human ability to create, learn, make decisions, and think, opening a new era in human-machine collaboration.
Deep Blue showed that it was capable of defeating one of the strongest chess players ever with a “programmed understanding of the game of chess”.
Will we ever be capable to make a similar strong cognitive chess program that learns and plays the game like humans do? Could it explain humans why it plays moves in human understandable terminology leading to improved human-machine collaboration in chess?
The center photo above was a clandestine operation I organized (4AM) to wheel Deep Blue from the Stage of the Equitable building to Rockefeller Center in hopes of getting some publicity on the Today Show. Deep Blue ended up spending a couple of minutes on air with meteorologist Janice Huff.
I recently read Luke Dormehl’s Thinking Machines and he had this to say on Deep Blue on Kasparov:
When Kasparov lost to IBM’s chess playing computer Deep Blue, he commented that he saw deep intelligence and creativity in the machine’s moves – hinting not that AI had developed these qualities, but rather IBM was somehow cheating by using human chess players behind the scenes.
With 20 years of Moore’s law and the pleasantly wide compute paths of neural network we now have self learning machines that use strategy in a human like way and reach international master level in 72 hours.
In my opinion IBM is not failing in the PR battle for AI despite being one of the pioneers.
When I go through various tech website there are talks about the activities related to AI of tiny/new companies and obviously the BIG ones however I hardly read anything about the prowess of IBM past and the current
developments related to IBM AI.
So I appreciate the fact that AI is able to help make decisions and would love to find some way to apply it to my day-to-day work. It seems, however, that AI is much more useful at a high (corporate) level for forecasts and data mining. At a low level (Dept level) it seems much harder to apply. Are there any ideas on how to apply it at a low level?
“…Each blind man feels a different part of the elephant body, but only one part, such as the side or the tusk. They then describe the elephant based on their partial experience and their descriptions are in complete disagreement on what an elephant is.” – The Rigveda This parable is helpful describing the problem that we […]
Headlines abound about doctors cowering from AI. The reality: not so much. I’m a physician, and don’t think that’s a realistic concern. Rather, I envision a future in which AI-enabled insights help health and medical experts deliver patient-centered, personalized, value-based care. The future is here. For those of us at IBM, we are augmenting experts’ […]
In the U.S. about 20 percent of adults suffer from a mental health condition, ranging from depression to bipolar disorder to schizophrenia, and about half of those with severe psychiatric disorders receive no treatment. While early identification, diagnosis, and treatment for patients with psychosis tends to mean improved outcomes, there continues to be significant barriers […]