Advances in fields like artificial intelligence and behavioral analytics are helping us keep pace with ever-more resourceful criminals.
There are a number of interesting ways advanced technology is being used to combat criminal activity, especially as more of that activity takes place digitally. Increased use of artificial intelligence and machine learning are two of the big ones – but the thing that really excites me is the use of behavioral analytics for identity verification.
When you think of all the peer-to-peer money movements that are taking place today, including transfers between people in different countries, there are a multitude of small-dollar money movements going on at any given point in time. Is that a genuine worker who is sending money back home to his family, or is that someone who is sending USD$100 or USD$200 to fund some sort of criminal or terrorist activity?
Determining identity is fundamentally risk management for transactions — somebody wants to do something over here, someone on the other side of the fence is trying to make a call on whether they should let that transaction take place. It’s difficult work, and increasingly law enforcement agencies, banks and companies must look at as many facets of the identity puzzle as possible to make those calls.
Essentially, identity centers around four components:
- Essence of who you are: Your DNA, how you look, your fingerprints, your biometrics. Things that are undeniable about you. This is perhaps the closest thing to truth about your identity.
- Legal documents: Your passport, driver’s license, social security card, birth certificate and other “official” documents.
- Electronic representation of yourself: How you choose to represent yourself on email, Facebook, Twitter, Instagram and other social media.
- Behavioral: Which spans an incredible range about the way you do things – anything from your purchasing patterns to the way you normally conduct business to even something as seemingly inconsequential as the way you hold your phone (more on that in a minute).
Some of these identification methods can no longer be counted on in isolation. While the essence of who you are can be confirmed through things like facial recognition and fingerprinting, the legal documents we so heavily rely upon today can be faked. In fact there are plenty of document verification services that have popped up to look at things like whether the font on the passport looks right, etc.
Likewise, electronic representations of people can be less than truthful. We all have a choice about how we want to present ourselves on social media, and many people create entirely false personas. In April, Facebook announced a crackdown on 30,000 fake profiles in France alone – that’s a big number.
Increasingly, behavioral analytics are coming into play to verify identity – or threats to it – by looking for differences in the way people do things.
How behavioral analytics work and how it will help
Take your smartphone, for example. Most people use a finger to unlock the phone with a PIN code. If I happen to get your pin code, I can technically break into your phone, but that may soon change. There are companies that are now looking at things like how you normally hold your phone, with which hand, and how much pressure you use when you tap on the screen. So although I may enter the right pin code, the system will be able to say, “Well, I’m not sure that’s really you because that’s not how you normally hold the phone. I have no record of you holding the phone that way” and deny access or ask for an additional verification method.
When you couple all these components together around this notion of identity, you can use some really interesting and advanced technology components to try to solve some problems…from the above “Is this person holding the phone the rightful owner or not?” to “Does this credit card transaction look normal?” to “These transactions are occurring at a time of day I would not have expected.” Credit card companies are already making good use of behavioral analytics by flagging purchases that seem unusual based on a user’s location and typical buying patterns.
I think going further down this behavioral identity road is both fascinating and compelling. Here at Thomson Reuters, we already have a number of really valuable data sets and propositions to help reduce the risk of fraudulent, criminal and terrorist activity through identity verification. Our CLEAR solution, as one example, was critical to identifying the perpetrators in the 2015 San Bernardino shootings, which enabled law enforcement to catch them and prevent further casualties. Our wide range of assets gives us the ability to aggregate from multiple places, better streamline the onboarding process and do the verification/risk management piece that fundamentally is what identity is trying to solve. And as more advanced behavioral analytics and algorithms come into the mix, these solutions will become even stronger.
The World-Check database is used around the world to help to identify and manage financial, regulatory and reputational risk.
CLEAR provides unparalleled insight to investigative and fact-finding tasks.