Report: AI ‘Friend and Foe’ in Financial Crime

Artificial intelligence (AI) continues to emerge as both a friend and foe in the fight against financial crime, a new report by software company ComplyAdvantage finds.

“The State of Financial Crime 2024” identifies the criminal use of AI as an emerging fraud challenge while revealing that most financial institutions are investing in technology to combat this growing threat.

“On the one hand, AI is being deployed by criminals to perpetrate fraud, launch attacks against individuals and corporations, and gain access to the international financial system,” the 107-page report notes. “At the same time, AI is increasingly being deployed in solutions such as customer onboarding, adverse media and sanctions screening, transaction monitoring, and automated reporting to regulators.”

The report finds that two-thirds of financial industry respondents think the use of AI ­– such as deepfakes, sophisticated cyberattacks, and the use of generative AI to create malware – by fraudsters and other criminals poses a growing cybersecurity threat.

Because of this, banks and other financial institutions are increasing their defenses against these threats, with 86 percent of respondents saying their company is investing in new technologies.

Iain Armstrong, a regulatory affairs practice lead at ComplyAdvantage, said, “Firms should ensure that they remain abreast of guidance and international agreements around AI. Compliance leaders exploring AI solutions should ensure they understand and document the rationale behind exploring a solution and have the right clean data to train AI models. Documentation proving that they understand how the AI models work is also essential.”

The report finds that 68 percent of respondents say they have a good understanding of how legislators and regulators plan to regulate AI technologies, and more than half say they are well prepared to meet proposed AI legislation.

Despite financial firms’ increased use of AI, the report notes that a majority of consumers remain uncomfortable with AI, even when it is being used to protect them. However, nearly half – 47 percent – of financial industry respondents said they do not prioritize explaining their use of AI to their customers.

On a global basis, the AI market was anticipated to reach $241.8 billion by the end of 2023, the report says.

In areas where concerns remain, like explainability, regulators increasingly look to those using or providing AI models to have clear and understandable information on the AI model’s capabilities and limitations and transparent and traceable decision-making processes.

With a wide range of legislation likely to be implemented in 2024, firms will also likely receive greater clarity on policymakers’ expectations, the report adds.