Banks navigate trust hurdles as AI investments grow

Banks are rushing to adopt AI, but many are struggling to make it work.

A new global study by SAS and IDC, surveying over 2,300 executives, found that nearly one in four banks operate at the highest level of “trustworthy AI” maturity. Yet almost half are stuck in what authors call the “trust dilemma,” where confidence in AI doesn’t match the ability to implement it effectively.

“The trust dilemma illustrates the difference between perception and practice: the trust in AI’s promise versus the organizational capacity to ensure its reliability,” said Chris Marshall, vice president at IDC.

It represents a major risk in banking, where functions like fraud prevention, credit risk and compliance demand explainability and reliability. 

“Misalignment here isn’t just a tech issue, it’s a regulatory one with high levels of reputational risk,” Stu Bradley, senior vice president of risk, fraud and compliance solutions at SAS, told The FR. He noted that banks can bridge the gap by strengthening governance, centralizing data stewardship, and ensuring risk and innovation teams work together so trust moves from principle to practice.

Banks were more likely than any other sector to rate themselves as highly mature, with 23.4% reporting they operate at the highest level of trustworthy AI maturity, meaning they have embedded governance, explainability, compliance and ethical safeguards into their AI programs.

Despite strong enthusiasm for AI, the study found that many banks face significant hurdles putting it into practice. Nearly half (47%) either hesitate to deploy reliable systems or rely too much on ones that aren’t ready.

Barriers to execution

Banks struggle with data governance gaps (44.6%), a shortage of specialized AI talent (42.3%), and fragmented, non-centralized data (40.7%). Therefore, despite strong interest in scaling AI, foundational elements including clean, connected data and skilled people often aren’t in place.

More than half of banks (51.6%) are building core AI platforms to address these challenges. Just over four in 10 (42.9%) are creating dedicated data science teams, and nearly 39% are investing in training and reskilling existing staff. The emphasis is on laying the technical and organizational groundwork.

While banks are leaning into AI as a source of growth and differentiation, their ability to capture that value hinges on making the infrastructure work.

“Stronger infrastructure and compliance frameworks improve the trust dilemma, but globally, many organizations must still invest in data strategy before scaling transformative AI," said Neil Ward-Dutton, vice president of AI, Automation, Data & Analytics Europe at IDC.

Despite these hurdles, banks continue to double down on AI investments. More than two-thirds expect to boost spending over the next year. While 11.5% plan to raise their AI budgets by more than 20%, nearly 6 in 10 (57.2%) anticipate increases of up to 20%, making banking one of the most aggressive sectors in the study.

The focus of that investment is clear: banks see AI as a growth driver, not just a tool for cost-cutting. Nearly two-thirds of leaders (63.4%) cite product and service innovation as AI’s top value, followed by process efficiency (60.9%) and workforce productivity (51%).

Insurance shows modest progress

The study found that 20.1% of insurers operate at the highest level of trustworthy AI maturity, trailing banks but still slightly above the global average of 19.8%. Insurers also face the trust dilemma, with 43% caught in underutilization or overreliance — a smaller share than banks’ 47%.

Insurers’ posture is more cautious. Their challenges echo banking — weak data foundations, governance issues, and talent shortages — but they are less likely than banks to prioritize architectural solutions to address these issues. Instead, they lean more on skills development, with 52% focused on training and reskilling.

Compared to other industries, banking, insurance and life sciences still stand out as “focus industries” where AI adoption happens alongside regulatory scrutiny and broad societal impact.

“In banking and insurance, AI reshapes fraud detection, risk management and personalization, where sensitive data and strict compliance make governance essential,” the study noted.

Bridging the enthusiasm and trust gap

Banks and insurers can’t rely on spending alone to unlock AI’s potential. The real differentiator will be whether institutions can embed governance, transparency, and reliability into every layer of their AI strategies, the study found.

“Trust is the cornerstone of successful AI adoption,” said Chris Tobias, general manager at Intel Americas, technology leadership and platform ISVs. “Building transparent, secure, reliable, and ethical AI systems ensures that organizations can confidently leverage AI to drive innovation and achieve transformative outcomes.”