Is artificial intelligence enough to win Jeopardy? It was elementary for Watson, IBM’s super computer designed to take on former champions, Ken Jennings and Brad Rutter, this week. He won by a lot – more than $50,000 separated him from the humans.
With his four terabyte library and ability to complete multiple processes at once, Watson’s data analytics and technology were impressive. However, the computational research lacked a human element, which we feel is the key to data analytics.
As Casey Johnston reported in ARS Technica on Tuesday, Watson lacked an understanding of the nuances in the human language and the ability to hear what is happening around him. He has to be fed the information in text form, which he processes at the same time as the other contestants. He is not hooked to the Internet, so running the answer through his databases is very similar to that of the human response.
Watson, Can You Hear Me?
The difference in the human intellect and machine capabilities were apparent starting with a “Name that Decade” question, where Jennings incorrectly answered, “What is the 1920s?” Watson produced his own “What is the 1920s?” response.
Another area where Watson goofed on language was in the answering of a question on Olympic oddities. When Jennings incorrectly answered that Olympic gymnast George Eysner was missing a hand, Watson chimed in with a “What is a leg?” He missed the word “missing,” which caused the segment to be reshot and his answer was deemed incorrect.
Artificial Intelligence – No Knowledge of Nuance
It’s the nuance of details like this that computational research can’t uncover. So even though Watson did a great job at cutting through the information overload to find insights … even if he processed 200 million pages of content to uncover this insight … the human element remains the key to business intelligence. And if this human element in BI interests you, learn how Spotfire, among the easiest to use analytic applications, can help you spot trends as fast as Watson thinks.
Watson – A Key to More Intelligent Data Interactions with Humans
It’s inspiring to see how Watson’s technology will be used to create a “cybernetic” physician assistant through a partnership with Columbia University and the University of Maryland.
And, Watson’s real promise is his technology’s enhancement to the experts. Spotfire’s Brad Hopper recently discussed the importance of the human element in BI on Data Management Radio, when he said, “a spreadsheet doesn’t understand your business.” But, empowering humans with powerful data analytics tools like Watson can help humans understand their businesses and make better decisions faster.
