Think Tank

A.I. for Compliance: An ‘Artificial’ Answer to a Very Real Problem

When it comes right down to it, are all these corporate scandals the result of a lack of intelligence? Specifically, of the artificial variety?

Businesses today are getting caught out for any number of violations relating to regulation, safety, security and immoral behavior. It’s more than any fallible human can manage. So it stands to reason that artificial intelligence can help.

The old-fashioned way of ensuring compliance — with armies of individuals spread across functions and locations, who might never speak with one another — is clearly broken. Businesses are spending millions on audits, undergoing endless testing, interviewing, document collection and system reviews, only to trip up time and again. Now add to the burden of compliance the huge costs associated with supply-chain disruptions, and long-lasting damage to brands.

Can A.I. come to the rescue? It has managed to seep into just about every other aspect of corporate operations. Meanwhile, the traditional compliance function is like “looking for a needle in a stack of needles,” says Beth Haddock, chief executive officer and founder of Warburton Advisers. Perhaps A.I. can help companies to head off the critical lapses that lead to a serious compliance breach.

Any discussion about the application of A.I. to a particular task must be preceded by a definition of the term. Does it represent an attempt to duplicate the workings of the human brain, or exceed it? Does it work through brute force — like instantly calculating the millions of possible moves in a game of chess — or is it truly “creative,” smoking out solutions in mysterious and non-linear ways?

These are questions that A.I. researchers have yet to fully answer. In the meantime, though, while the debate over A.I. rages on, why not apply it to a complex area such as compliance?

One place where A.I. excels is in the management of data. There’s just too much information out there today, moving too quickly, for humans to make sense of. “Businesses are too global and too complicated,” says Haddock. When it comes to compliance, “with humans you’re behind, always trying to catch up.”

Information these days comes in three basic forms: open data, in the form of studies, papers and analyses, often available on the internet; public data, usually free but protected by copyright and requiring adherence to some form of legal agreement; and private data, available only through subscription or permissioned access.

Businesses are constantly struggling to toggle between these categories of data, unable to tell which kernels of information are relevant and which should be ignored. That’s where A.I. comes in, with its ability to sort through the “noise” and — guided by criteria established up front by human programmers — select which issues companies need to be paying attention to.

It’s not a question of sweeping the humans aside. The effective use of A.I. for compliance purposes requires cooperation between those who are running the system and experts who have a deep understanding of laws, regulations, industry fundamentals and enterprise risk.

Haddock likens the governing structure to a sandwich, with the technology layer residing between a compliance officer who sets the rules at the outset, and another who can act on the analytics performed by the A.I. engine.

Haddock says A.I. has grown to become an increasingly important tool for compliance officers in the post 9/11 years. It has been applied to such tasks as detecting terrorist activity, money laundering and the tracking of financial transactions around the globe.

A.I. has come to perform much of the grunt work that used to burden human analysts, such as collecting information, entering data and filling out forms. Now, it’s poised to attack much more complex tasks, such as know your customer (KYC) compliance. A.I.-driven systems can help to assess the risks involved in doing business with particular companies and individuals, based on data that might not be available or visible to an unaided human compliance officer.

Haddock says A.I. is growing rapidly in sophistication, raising the question of whether compliance departments can keep up with developments in the technology. At the same time, the role of those individuals is changing. No longer mere amassers of data, they’re required to be high-level strategists, taking a broad view of the company’s compliance program.

In the future, says Haddock, A.I. won’t just offer protection from problems that happen to arise in the course of doing business. “You’ll be able to spot issues earlier, and be more productive. The role of humans will only increase. A.I. will allow compliance to do more.”

Comment on this article

No entries

Nothing found in the guestbook.

CAPTCHA image for SPAM prevention If you can't read the word, click here.