IBM's Watson to Tackle Computer Security, Threats

IBM's artificial intelligence system Watson is to lend its hand to security. It will help scour millions of reports to try to learn more about current online threats.

The Watson system was originally designed to see if a computer could compete on Jeopardy!, the show where contestants are given an answer and have to try to figure out the question. It was considered a challenge for computers because, although they can store a lot of information and process it quickly, they don't always have the human ability to decipher meaning and context.

In the event, Watson prevailed against two human champions on a special episode of the show. The technology has been made available for public use, the idea being to combine the reasoning skills of humans with the speed of computers. One suggested example was for Watson to build up a database of medical research and then look at a set of symptoms and try to figure out possible conditions that cause them. The same has now been adopted for computer security.

App Will Use Natural Language

The new project is an IBM app for businesses, along with an associated voice recognition system that will allow IT staff to ask questions in natural language. The project involves Watson analyzing the contents of more than a million security documents to learn more about security threats, causes and solutions. (Source: wired.com)

The theory is that this bridges the gap between human and computer solutions. On the one hand, humans can only read so many documents in a given time and then only recall a certain amount of the data. On the other hand, it's not easy for a computer to search through security reports written by humans because the terminology and phrasing will vary significantly depending on who wrote the report and for what purpose. The same applies to the way that IT staff phrase questions about what could be causing a security threat.

Threats Could Be Solved In Minutes

Watson aims to read the documents and turn them into a searchable database of clues about the wide variety of security threats, including viruses which replicate and mutate to make it harder for antivirus software to keep up and detect them.

More than 40 businesses have now signed up to use the service and IBM believes it could reduce the time it takes customers to diagnose the cause of a security problem and the most appropriate solution from days or even weeks to a matter of minutes. (Source: ibm.com)

What's Your Opinion?

Are you impressed by this approach to tackling security threats? Can you see any drawbacks? What other areas of life do you think human-style reasoning with computer speeds could be helpful?