Banks occupy a unique position when it comes to intercepting crimes, such as drug and human trafficking. Criminals involved in these types of illicit operations often move the proceeds from trafficked drugs or victims through banks, hiding the source of their funds and decreasing detection from law enforcement.

In cases when banks manage to detect suspicious activity indicative of money laundering, they can provide leads and evidence law enforcement can then use to discover and prosecute those involved.

In the past few years, artificial intelligence has emerged as a solution to help banks zero in on transactions that herald criminal activity. Machine learning allows for the identification of suspicious patterns, and AI systems are able to comb through large amounts of data more precisely and efficiently, catching instances of money laundering faster and with more accuracy than ever before.

“What the AI solution does is integrates all the data from the different bank’s siloed systems, and comes back with a good answer on where the criminal activity is,” explains Dan Stitt, director of financial crime analysis for QuantaVerse, a technology company that specializes in risk reduction for banks, including through AI and machine learning techniques. Stitt also has over 25 years of federal law enforcement experience as a former agent of the Drug Enforcement Administration and former assistant inspector general for investigations at the Export-Import Bank of the United States.

“(During) my investigative career with the government, a lot of my leads were from financial institutions,” Stitt told Forensic Magazine. He explained that law enforcement at the local, state and federal level can use data on suspicious activity provided by banks to the Financial Crimes Enforcement Network (FinCEN) database. Banks also call in phone tips regarding possible criminals involved in drug trafficking or other illicit operations.

“In the banking compliance world we’ve actually discovered human trafficking cells and human slavery situations, reported that to law enforcement, and they made arrests within a 3-month period,” Stitt said.

Possible human trafficking situations have been a priority for banks, according to Stitt. In 2014, FinCEN released an advisory on how financial institutions can catch “red flags” of human trafficking and smuggling within financial transactions and report the activity to authorities. In March 2017, the Royal United Services Institute released a report titled “Disrupting Human Trafficking: The Role of Financial Institutions,” which focused on the role of banks in both the U.S. and the U.K. in using their unique position to intercept this type of crime, which affects 20.9 million people globally, according to the report.

“The increased focus on human trafficking by financial institutions has resulted in some success. According to the Manhattan District Attorney, his office has, with the help of financial institutions, ‘been able to secure convictions against traffickers without having to rely solely on the testimony of victims who often suffer emotional, physical, or sexual abuse,’” the report states, also noting that, since the release of the FinCEN advisory, more banks have begun specifically identifying suspicious activity as possibly related to human trafficking in their reports.

AI can further strengthen banks’ ability to aid law enforcement by not only identifying more potential cases of criminal activity, but by giving investigators more to work with when it comes to details about the activity, Stitt says.

“What’s really useful to the government and to law enforcement is quality suspicious activity reports—drilling down on what’s really suspicious, where the criminal activity is taking place, where we can prevent and report crime to law enforcement promptly—that’s where we’re able to help with our AI-driven technology,” he explains.

Algorithms can identify suspicious activity more quickly than manual methods, including transactions with missing information, a large amount of transactions below $10,000, transactions made under slightly different names or addresses, an abnormal amount of transactions going to the same account and a series of transactions going to the same account over a long period of time, as outlined in a 2016 QuantaVerse report and case study.

According to the case study, implementing AI and risk reduction technology, including these key algorithms, at a major bank resulted in the identification of $6.2 million in previously unidentified laundered money in less than one week.

“As a result of this multivariate process, QuantaVerse found many more questionable transactions being run through the bank each month. The parties identified by this process included criminal individuals, shell corporations, and a law firm that created the identified shell corporations,” the report states.

These types of discoveries made by AI increase the ability of banks to be greater partners to the justice system, with the growing technology “assisting banks to be quicker by sharing information sooner with law enforcement” in an attempt to get ahead of some criminals, Stitt says.