One area where it's just getting started is law enforcement, where Watson's data-processing abilities have enormous potential to speed up investigations, which generate thousands of pages of paperwork. The computer can process those documents in the time it took to read this sentence, potentially giving detectives the best actionable leads much faster than before.
See also: Top 25 Free iPhone Apps of All TimeBut Watson could have an even greater impact in the way law enforcement itself is carried out. The police shooting of unarmed black teenager Michael Brown in Ferguson, Missouri, sparked national discussion about police overreaction, racial profiling and the militarization of law enforcement.
There was rancor and rhetoric, to be sure, but a lot of smart and informed people also ventured into the debate. Still, there's no one who could know all the relevant data to give the best insights on those topics.
The cognitive copNo one except Watson, that is. At an IBM event to showcase the fields where Watson is making a difference, Roberto Villaseñor, chief of police in Tuscon, Ariz., cited Ferguson as a good example where Watson can sort things out.
"When we have a significant event, we have a formal board of enquiry," Villaseñor explained. "We try and determine: Are we doing the best we can in that situation? But because we're human, that limits the amount of information we can gather.
Cognitive computing can come up with a lot more information to review, digest and possibly incorporate." Villaseñor is one of the backers of IBM's CopLink software on the police side, which is already helping with data analysis. He cited a specific investigation that's generated a mountain of paperwork: an abduction of a 6-year-old child in Tuscon right out of her home.Cognitive computing can come up with a lot more information to review, digest and possibly incorporate.
Counting just the documents related directly to the case — including police reports, statements, lab results and more — there are 15,000 pages to go through. When you factor in things such as outside agency reports that might be relevant, the backgrounds of people involved and information online, pages can easily get into the tens of thousands.
Villaseñor is hopeful that Watson and cognitive computing systems like it, which have just begun to touch the law-enforcement space, will be able to process all that information, and give investigators real leads faster.
"There may be something in lead no. 25 that doesn't make sense until you get to lead no. 2,050," Villaseñor said. "How is a human going to tie those things together? Cognitive computing can. But we still have to look at it. It cannot be a computer or a human analysis. It has to be and. We say, 'Watson said this — let's go check it out.'"
He also would like to see that computational power directed against the bigger police issues being debated right now over Ferguson and other flashpoints in law-enforcement. The situation in Ferguson has led to an at-times heated discussion about what's wrong with law enforcement, particularly in black urban communities, which Villaseñor said has led to a lot of misinformation spreading.
"There are a lot of theories being thrown out in the news media," he said. "Being able to trudge through all the information and data, and put out accurate information, as opposed to speculation or analysis based on speculation and supposed truth that's being put out through third-party hearsay ... you need to filter through that."
Big data for big solutionsVillaseñor is being a little unfair. Most of the commentators on Ferguson cite some statistics or studies to support their point of view. But the idea of cognitive computing would be to apply the computational power of an entity that can digest all of the relevant data. Watson could gobble up every relevant case, statistic and incident regarding law enforcement in minority communities in its effort to come up with solutions.
Take the issue of police militarization. There are records for what equipment was sold to which agencies and departments, how those tools are deployed and how they've been used. However, for a human to study the issue, they'd have to do most of the legwork of assembling the relevant data themselves, even with computers — a process that could take months.
Get Watson on the case, and it'll find the right stuff to look at no matter how deep it's buried — in considerably less time. But it's not effortless; programmers will need to be rigorous about how they define exactly what they want Watson to look for. But once that's done, finding all the relevant data, usually a tedious and lengthy process, will be the easy part.Get Watson on the case, and it'll find the right stuff to look at no matter how deep it's buried — in considerably less time.
"There are mounds of information out there that we're going to need help sorting through to help us not necessarily answer the question, but at least define the problem," Villaseñor said. "We need to get the data-driven information, and not go with anecdotal information because there's a lot of emotion behind it. We need to try and get past the emotion and find the truth. It may be bad, but we need to find out what it is so we can adjust."
Have something to add to this story? Share it in the comments.