Finance firms face surging AI risks as conduct incidents average USD 14 million

• The share of C-suite executives identifying AI-related conduct risks as a top material risk has jumped from 16% over the past three years to 56% over the next three years, ranking first among non-financial risks.
• While executives report that major business conduct risk incidents rose by 55% between 2023 and 2025, with each incident costing USD 14 million on average, the majority of firms continue to invest in conduct data reactively rather than preventively.
• As financial firms scale AI across risk workflows, trusted data becomes essential to keep outputs relevant, accurate, and auditable – helping prevent hallucinations from spreading across models, dashboards, portfolios, and decisions.

RepRisk, the world’s most respected DaaS company for business conduct risks, today released new analysis from its global Business Conduct Risk Intelligence Report 2026, based on a survey of more than 500 C-suite executives across banks, asset managers, asset owners, and other financial institutions, conducted in collaboration with Oxford Economics.

Major risk incidents rise as costs mount

RepRisk’s new analysis estimates that firms face USD 28 million to USD 43 million in annual cost exposure from reputational and business conduct risks, with companies experiencing two to three significant incidents per year on average and each incident costing around USD 14 million. The most severe incidents average USD 37.6 million, highlighting the financial value of earlier detection and prevention. Even modest improvements in monitoring – such as reducing incident frequency by 5% to 10% or accelerating escalation before issues intensify – could help mitigate multi-million-dollar losses annually. By enabling firms reduce blind spots, shorten decision cycles, strengthen governance, and improve auditability and decision confidence, structured business conduct risk intelligence creates measurable value, with C-suite executives surveyed expecting the ROI from these capabilities to double within three years. 

Beyond direct costs, major incidents can lead to the loss of key investors and clients, regulatory sanctions, and damage to brand reputation. Proactive business conduct risk mitigation is therefore becoming a value driver, helping firms protect performance, reduce blind spots, and act before risks escalate.

AI risks surge as adoption scales

The report found that only 16% of executives identified AI-related conduct risks as a top material risk over the past three years – but 56% expect them to be a top material risk over the next three years. Against this backdrop, executives report that major business conduct risk incidents rose 55% between 2023 and 2025, while 67% say overall risk complexity has increased over the past year.

AI risks are surging just as banks, asset managers, asset owners, and other financial institutions are embedding AI into core workflows, from transaction due diligence and risk monitoring to portfolio oversight, compliance, KYC, and stewardship engagement. For financial institutions, the stakes are magnified by scale. Business conduct risk data can inform decisions across large portfolios, client relationships, counterparties, transactions, and internal control frameworks. If flawed or inconsistent data enters AI-driven workflows, errors can spread across models, dashboards, portfolios, and decisions – becoming difficult to identify, explain, or reverse after the fact.

“As we celebrate RepRisk’s 20-year anniversary, our founding mission, to bring transparency to business conduct risks to drive positive change, has never been more relevant,” said Philipp Aeby, CEO and Co-Founder of RepRisk. He added, “AI is moving deeper into financial decision-making, but models are only as trustworthy as the data and guardrails behind them. Hallucinations, inconsistent sources, and opaque methodologies can quickly scale into costly business decisions. Financial leaders want the speed of AI without the risks of black-box automation. They want trusted, AI-powered risk intelligence with humans in the lead, built on data they can explain, defend, and stand behind.”

The survey points to a clear preference for human-led AI when business conduct risk data informs material decisions. Across the full sample, 73% of executives report using human-AI hybrid approaches, while 67% trust hybrid data for material risk and investment decisions, compared with 35% for AI-only approaches. This preference is even stronger among banks, where 74% of respondents express confidence in human-AI hybrid data, underscoring demand for technology enhanced by expert oversight, rather than full automation.