Looming Large: Artificial Intelligence’s Promise and Threat

Looming Large: Artificial Intelligence’s Promise and Threat

For anti-financial crime (AFC) professionals, the potential of artificial intelligence (AI) was one of the biggest themes of 2023, as we explored how to use vast amounts of financial intelligence to identify, interdict and seize the proceeds of crime and ultimately bring criminals to justice.

We pursued that theme in earnest in 2023, identifying the promise of machine learning and generative AI within these ACAMS Today pages, on ACAMS moneylaundering.com, and at ACAMS’ events, providing opportunities for AFC professionals to question what, exactly, AI is―and what it can realistically be expected to accomplish.

Notably, at several ACAMS Assembly events we heard regulators talk up AI as a new, proactive paradigm that could put law enforcement (LE) a step ahead of financial criminals. “I think we can all agree that the status quo is not the answer,” the chief executive of Canada’s financial intelligence unit, Sarah Paquet, said in a keynote at ACAMS’ Assembly Canada in mid-November.1

Challenging attendees to file high-priority suspicious activity reports in real time, the FINTRAC director vowed to meet that timeliness with accelerated processing of information to LE and feedback to the filing institutions.

The potential for AI to support “real-time reporting” which would allow LE officials to gather intelligence that profiles criminals―and stops them―in the act, seems to be the Holy Grail for AFC professionals. But as Paquet made clear, achieving anything close to real-time reporting of suspicious transaction reports (STRs) depends on the adoption of AI by both regulators and financial entities.

While her remarks were generally well-received by attendees, professionals at some of the big Canadian FIs questioned how true can “real time” be, and shared concerns about both whether AI and its tools will continue to mature and if it can possibly live up to its potential.

“It’s a great opportunity, but the question will be: ‘[I]s my institution’s data structure ready to take advantage of it?” Michael Donovan, chief anti-money officer at CIBC, noted during a later panel discussion, as reported in ACAMS moneylaundering.com.2 On a related note, find “Open Banking and Financial Crime Risks in Canada” in this edition.

I am reminded of Willie Sutton’s famous reply to the question of why he robbed banks (“that’s where the money is”). Today, of course, the money is online. And while we know criminals are still relying on old-fashioned methods of fraud to commit financial crime, we also know they are turning to increasingly sophisticated stratagems including new technology and even AI tools to steal from hapless individuals, financial firms and other businesses.

Limiting their impact means not only having the capacity to identify transactional crime in real time but being a step ahead of the individual scammers, transnational criminals and nation states now using AI with devastating results for their victims.

For instance, criminals are already capturing brief voice recordings of unsuspecting individuals through a variety of ruses, including wrong numbers, and using AI applications to imitate their victims’ voices to commit fraud, according to Chris Lynam, director general of the National Cybercrime Coordination Centre and the Canadian Anti-Fraud Centre of the Royal Canadian Mounted Police.

These AI-generated deepfakes can be effective enough to deceive family and friends of the victim. They can also fool banks and other financial firms that utilize investment voiceprint verification systems, Lynam told Assembly Canada attendees.

At the Assembly in Las Vegas, the month before the Assembly Canada, the FBI’s Chad Linnerooth anticipated Lynam’s remarks when he told the ACAMS audience that the use of AI and deepfake technology by criminals will grow, fueling a whole new generation of scams, making fraud harder than ever to detect. For further discussion on this topic, I refer you to “Exploring the Role of Generative AI in Enhancing Financial Crime Compliance,” in the September-November issue of ACAMS Today and “AI and the FDIC’s Disaster-related Regulatory Adjustment” in this issue.

Even as we work to leverage these new tools to fight crime, a broader debate with frightening sci-fi elements is taking place: Is there a danger that the AI entities we are counting on to serve us could go rogue? Think Isaac Asimov’s I Robot.

But AI, for good or for ill, is clearly here. Its potential uses were an AFC theme of 2023; the need to work together to master it for good is likely to be the task of 2024.

Kieran Beer, CAMS
Chief Analyst, Director of Editorial Content
Follow me on : @KieranBeer
“Financial Crime Matters with Kieran Beer”

  1. Fred Williams, “Toronto Reporter’s Notebook,” ACAMS moneylaundering.com, November 20, 2023, https://www.moneylaundering.com/news/toronto-reporters-notebook/
  2. Ibid.

Leave a Reply