Financial leaders warn against giving AI 51% stake in decision-making
Financial institutions have long regarded customer data as proprietary assets, but the new PYMNTS report says: “With fraud and find trust in the bank’s data flood”, suggesting that a greater competitive advantage may come from knowing when information is aggregated.
In the digital economy, fraud to this week's development, government data sets are shrinking, and banks and credit unions have found that cooperation, not just competition, can make the difference between trust gain and loss.
The study is part of the “Find a Reliable Signal in the New Data Reality of Banks” series, which examines how banks, credit unions and fintech rethink data strategies as artificial intelligence (AI) plays a bigger role.
The report draws on the views of executives from Velera, Entersekt and Concora Credit, who warn that financial crime now requires a “team movement” approach.
Their shared information: data is still essential, but its reliability depends on balance – combining history with real-time signals, human supervision with machine intelligence, and competition with institutions that share intelligence.
- Fraud is forcing speed into the system. Pradheep Sampath of Entersekt said traditional government feed (from Fed fraud reports to Fincen’s documents) was too slow to deal with today’s threats. Almost all respondents in the series pointed out that the need to combine historical bureau data with behavioral analysis, device fingerprinting and geolocation markers that may detect abnormalities at this moment.
- Collaboration is in quarantine. Velera's Jeremiah Lotz highlights how a consortium model that aggregates thousands of credit union data has produced stronger defense capabilities. 100% of executives surveyed described fraud management as “competitive neutral”, one of the few areas where competitors can safely share signals without ceding their advantage. Encrypted and federated data models (such as encrypted and federated data models) are called bridges.
- Alternative data is gaining the foundation. Kyle Becker of Concora Credit reports that stratification of about 12 alternative datasets each year can improve underwriting and fraud detection. In particular, cash flow underwriting, as a tool, can expand credit access and strengthen defense capabilities. The report notes that institutions even find “one or two” of viable new data sources each year that can even exacerbate improvements in risk management over time.
What emerges from these findings is a shift in mindset. Rather than just treating fraud prevention as an arms race for better models and faster alerts, it is an ecosystem challenge for financial institutions to rise or fall together.
Advertisement: Scroll Continue
The executives interviewed by PYMNTS stressed that there is no dataset, no single tool, and no single company can completely resolve fraud. Instead, they call for “hierarchical intelligence” – a mix of bureau records, first-party transaction history, commercial data sets and real-time signals, and deployed through governance.
In a landscape where consumer expectations for safety and convenience continue to climb, the line between fraud defense and customer trust is blurred. As the report concludes, no reliable signal was found, but was built through cooperation, governance and willingness to see data as wall assets but as a common defense mechanism.
