Why Are Companies Slow to Adopt AI in Fighting Financial Crime?

Despite AI’s promise in fighting financial crime, firms remain cautious – citing regulation, cost, and data silos as barriers to adoption



Artificial intelligence (AI) and machine learning (ML) are becoming useful tools in combatting financial crime, but many organisations are hesitant to embrace these new technologies. A recent survey conducted by SAS in partnership with KPMG underscores this slow adoption, revealing that only 18% of financial institutions have fully implemented AI/ML in their anti-money laundering (AML) processes. A further 18% are piloting the technology, while 40% have no current plans to integrate AI/ML at all.

So, what’s holding companies back? Here are the key reasons behind the sluggish adoption of AI-driven financial crime prevention:

1. Regulatory Uncertainty

Regulatory bodies appear to be sending mixed signals when it comes to AI adoption. The survey found that only 51% of respondents felt regulators encouraged AI/ML innovation – down 15% from 2021. At the same time, the proportion of those who believe regulators are “resistant to change” has more than doubled to 13%.

“Regulators’ caution is understandable, but firms that hesitate may find themselves lagging behind,” says Jo Whalley, Director at Bigspark. “Those who proactively integrate AI with governance in mind will gain a competitive edge.”

2. Budget Constraints

While budget concerns have slightly eased (dropping from 39% in 2021 to 34% in the latest survey), they remain a major roadblock. AI adoption isn’t just about purchasing software – it requires investment in infrastructure, skilled personnel, and ongoing model training.

“AI isn’t a magic fix – it needs strong data management foundations,” notes Jo. “Without that, even the most advanced AI solutions won’t be effective.”

3. No Regulatory Mandate

Unlike some other compliance measures, AI-driven fraud detection is not yet a regulatory requirement. The survey found that 37% of respondents cited this as a key reason for slow adoption. Many firms are hesitant to invest in AI unless they are compelled to do so.

“Without clear regulatory pressure, many firms deprioritise AI adoption in favour of more immediate compliance concerns,” says Jo. “This short-term thinking could leave them vulnerable as financial crime tactics evolve.”

4. False Positives and Model Transparency

One of AI’s biggest promises in AML is reducing false positives. While 38% of AML specialists now see this as AI’s primary benefit (up from 30% in 2021), firms remain wary of the potential for unintended consequences.

“There’s a balancing act between catching real threats and not overwhelming investigators with false alarms,” says Jo. “AI models must be explainable to ensure trust and regulatory compliance.”

5. Skills Gaps Are Still a Challenge

The good news? The skills gap is becoming less of a concern, with only 11% of respondents citing it as a major challenge. However, some firms still struggle to find the right talent to develop, implement, and maintain AI-driven AML solutions.

“The talent shortage isn’t as severe as before, but expertise in AI-driven compliance is still a niche area,” Jo confirms. “Firms need to invest in training their teams to bridge the gap and get the most out of these technologies.”

6. Generative AI (GenAI) Raises More Questions Than Answers

Interest in GenAI is high – 45% of firms are either piloting or exploring its use in financial crime prevention. However, 55% have no plans to integrate it, indicating widespread uncertainty about its reliability and security.

“GenAI is still in its early days for AML,” Jo adds. “Firms need to test its capabilities carefully before widespread adoption.”

7. Siloed Data and Lack of Integration

AI’s effectiveness depends on high-quality, integrated data. The survey revealed that while 86% of respondents are integrating AML, fraud, and information security data in some way, only a third have a fully integrated case management system.

“The key to unlocking AI’s potential is breaking down data silos,” says Jo. “Firms that prioritise integration will see the greatest benefits.”

While AI holds immense potential for fighting financial crime, adoption remains slow due to regulatory uncertainty, budget constraints, and data integration challenges. However, firms that take proactive steps – investing in data infrastructure, aligning with regulators, and carefully piloting AI solutions – will be best positioned to stay ahead of evolving financial crime threats.

To find out more about our Fraud and Fincrime practice click here