Banking Tech Agenda Grapples With AI Security and Tokenized Payments
This week’s banking technology outlook focuses on how AI-driven security reviews can expose large remediation backlogs, citing Bank of America’s effort to fix software issues found by an Anthropic model. It also tracks payments modernization through 2030 and the growing use of stablecoin infrastructure by card networks and processors. The piece argues that banks should budget for AI risk remediation and treat tokenized fiat as part of mainstream money movement.
Bank of America reportedly assigned 1,000 contractors and 3,000 internal employees to address software defects identified by Anthropic’s Mythos model. The episode illustrates how model-based code and configuration checks can surface remediation work beyond normal engineering capacity, making AI security findings a planning and funding issue.
A Payments 2030 report from Finextra and ACI Worldwide frames the next four years as a chance for banks to gain value from earlier real-time, ISO 20022, and cloud projects. Separately, Mastercard’s BVNK integration and Marqeta–Coinbase card efforts indicate that tokenized fiat is increasingly treated as a normal way to fund payments.
These developments could affect bank customers, merchants, payment processors, and engineering staff. AI-driven security reviews may expose defects that require costly fixes, potentially influencing service reliability and technology budgets. Tokenized fiat and stablecoin rails may change how cross-border and card payments move, possibly altering fees, settlement speed, and access for businesses and consumers. Banks’ compliance and liquidity roles may expand, while regulators and auditors could demand clearer records of automated decisions. The overall impact may depend on how institutions balance remediation costs, innovation, and consumer protection.