As AI systems move into strategic decision making, banks face rising governance and accountability challenges. Gerry Kounadis of the National Bank of Greece explains why boards must retain control, strengthen oversight of vendors, and embed AI risks into
As artificial intelligence evolves from a supporting analytical tool into a system capable of shaping strategic decisions, financial institutions are facing a profound shift in governance and accountability.
Boards and senior executives are being forced to reassess how responsibility is defined in an environment where decision making is increasingly influenced by opaque and complex models.
According to Gerry Kounadis, Head of Group Data Privacy, Technology and ESG Compliance Advisory at National Bank of Greece, the core principle remains unchanged despite the technological shift.
“Ultimate accountability must remain with the board and senior management, because fiduciary duties of oversight and judgment are non delegable and cannot be transferred to an AI vendor,” he explains.
While vendors play a critical role in developing and maintaining AI systems, Kounadis notes that responsibility cannot simply be outsourced.
Emerging regulatory frameworks are reinforcing this view, even as they introduce new layers of accountability for technology providers.
“AI systems are treated as products, so providers may face strict liability for defects, non compliance, or security failures,” he says.
“The company bears primary responsibility for deployment decisions, while the vendor carries significant responsibility for the model's integrity and compliance.”
This evolving division of responsibility is particularly important as firms become increasingly reliant on third party AI providers.
In many cases, organizations depend on proprietary models that operate as black boxes, creating significant transparency and concentration risks.
Kounadis argues that such dependencies must be treated as critical risks within governance frameworks.
“Heavy reliance on a single proprietary AI model should be treated as a critical concentration and third party dependency risk,” he says.
To mitigate these risks, boards must implement stronger oversight mechanisms, including independent validation, contractual audit rights, and robust decision tracking.
Just as importantly, institutions must ensure they retain the ability to replace or disengage from a model if necessary.
“The governing principle should be simple - if an organisation cannot explain, challenge, or replace a model, it should not rely on it for strategic decision making,” Kounadis states.
Beyond vendor risk, the rise of AI driven decision making raises fundamental questions about the role of human judgment.
Regulators and courts are increasingly focused on where the boundary lies between legitimate decision support and the inappropriate delegation of managerial responsibility.
“The line should be drawn where AI stops informing managerial judgment and starts replacing it in practice,” he says.
