Banks outpace regulators on AI adoption, report finds

Financial regulators are trailing banks and different fiscal firms successful adopting artificial intelligence, adjacent arsenic concerns equine implicit however the exertion tin beryllium monitored safely, according to a caller study from the Cambridge Centre for Alternative Finance.

The survey recovered that much than 80% of fiscal services firms are utilizing AI to immoderate extent. It besides said 52% are already experimenting with agentic AI.

By contrast, regulators stay astatine an earlier stage.

Of the 130 regulatory authorities surveyed, 48% said they are inactive exploring AI adoption oregon are not engaged with AI astatine all.

The spread comes arsenic oversight challenges go much pressing.

Recent reports that Anthropic’s Mythos exemplary is often much susceptible than humans successful hacking person added to questions implicit whether manual oversight of AI usage successful fiscal services is workable.

The study said bundle engineering is presently the astir mature AI exertion successful finance. At the aforesaid time, it described this country arsenic a large transmission for cyber hazard transmission.

Among respondents, 48% identified adversarial AI arsenic a starring concern.

The survey besides highlighted a disagreement successful however risks are viewed.

AI vendors were little apt than fiscal firms and regulators to springiness precedence to adversarial AI threats and cyber and operational resilience.

“Further complicating this occupation abstraction is simply a notable cognition gap: AI vendors spot little precedence than manufacture and regulators connected some adversarial AI threats (35% versus 50% industry, 57% regulators) and cyber/operational resilience (32% versus 46% industry, 59% regulators),” says the 2026 Global AI successful Financial Services Report: Adoption, Impact and Risks.

Across each stakeholder groups, information privateness and extortion ranked arsenic the apical perceived AI risk, cited by 73% of respondents.

“These intersecting vulnerabilities tin besides provender into the apical perceived hazard crossed each stakeholders – information privateness and extortion (73% of respondents) arsenic delicate information is typically the superior people for the cyber exploits these vulnerabilities enable,” the study adds.

Other risks identified successful the study see exemplary hallucinations, unreliable outputs, opacity, constricted explainability and marketplace abuse.

Despite the accelerated gait of adoption, the study said the effects truthful acold person been concentrated connected ratio alternatively than concern exemplary transformation.

It besides recovered that fintechs are up of incumbent firms successful utilizing AI for lawsuit support.

Meanwhile, 76% of respondents astatine ample fiscal institutions said they find it hard to measurement the worth of AI deployment.

Most organisations surveyed said they are gathering connected outer AI models alternatively than grooming their ain from scratch.