53% Of Banking Leaders Say AI Is Outpacing Governance

New research commissioned by ABBYY reveals that banks and financial services organizations are investing heavily in AI, but more than half (53%) say it’s being adopted faster than they can govern it.

The ABBYY Who Answers for AI: The Governance Gap report, conducted by Opinium Research surveyed 1,200 senior managers in businesses of 100+ employees located in the US, UK, France, Germany, Australia and Singapore.

Concerns around data quality are the biggest barrier to achieving greater ROI from AI, with over a fifth of banking and financial services leaders (22%) citing it as their top challenge. Other key barriers include the cost of implementation (16%) and integration with existing systems (15%).

A huge 83% of banking/financial services leaders also reported that data governance requirements slow AI deployment. This could be impacting ROI, with just 41% saying their AI initiatives have exceeded expectations.

While 53% say AI is being adopted faster than they can govern it, the research revealed that effective governance can play a key role in helping organizations realize greater value from AI. Four in five (80%) banking and financial services leaders surveyed have an AI governance framework in place, and 79% of those say it has made their AI initiatives more successful – enabling easier adoption among employees, as well as easier deployment and scaling.

However, governance maturity varies across organizations. Only around half (53%) of financial services companies have an AI incident reporting process in place. The industry is somewhat more advanced when it comes to the ability to shut down AI systems: 41% of respondents say they have a rollback facility in place, compared with 35% across all industries while 39% say they have an AI kill switch, also above the 35% average.

Human oversight remains important, with 56% of banking and financial services leaders saying humans review all important AI decisions before action is taken.

Knowledge about accountability decreases the more junior employees are. Banking and financial services leaders at companies using AI say 88% of senior leadership knows who is responsible for developing, implementing, and managing AI across their organizations, but only 64% say the same for junior employees.

“Businesses are embracing AI, but our research shows that many are still struggling with key questions around accountability, governance and ROI,” says Roman Kilun, Chief Compliance Officer at ABBYY. “Senior management needs to do a better job of communicating what their responsible AI policies are company wide. In addition, AI requires organizations to know where data comes from, whether it can be trusted, how it can be used, and where it is stored and processed. Clear governance frameworks and a shared understanding of responsibility are critical to building confidence in AI.”

Banks and financial services organizations are continuing to increase their investment in AI, according to the study. Respondents said they expect investment to increase by a further 19.5% next year, slightly above the global average of 18.8%. The most popular tools being used among financial services leaders was Generative AI: 84%, Agentic AI: 60% and Purpose-built AI: 56%.

Worryingly, trust levels among financial services leaders are among the lowest across industries. Only 16% said they completely trust AI to operate without introducing unacceptable risks – compared with an average of 22% globally. In addition, just over a quarter (27%) of financial services leaders completely trust AI to produce accurate results, while merely 26% completely trust AI to comply with applicable regulations.

In addition, banking and financial services leaders are divided over who should be responsible when AI systems produce harmful or incorrect outputs. The survey found that nearly a third (31%) believe responsibility for harmful or false outputs should be shared between the organization and the vendor. Meanwhile a quarter (26%) think the organization using the AI should bear primary full responsibility, and less than a fifth (18%) believe it lies totally with the AI vendor. The rest said the end user or regulators should take the lead.