Jim Tarantino

In the context of evolving technological landscapes, the integration of artificial intelligence (AI) presents both opportunities and challenges for financial institutions. Before implementing AI across their operations, financial institutions need to establish key foundational elements to ensure successful AI adoption and risk mitigation.

Having a clear AI governance framework, model risk management and centralized standards in place is critical for enabling banks and credit unions to succeed in their adoption of AI tools and technologies.

1. Governance framework: The adoption of AI technologies requires a well-structured AI governance framework that comprehensively addresses the unique risks and regulatory considerations associated with these advanced technologies. Financial institutions should start with exploratory projects, such as proof of concepts, to gain insights into the operational and risk implications of AI. These insights can then guide the development of an AI governance framework that may either stand as an independent initiative or integrate into existing governance initiatives such as model or IT governance.

A financial institution’s AI governance framework should draw upon elements from established industry standards and regulatory guidelines to create an approach that aligns with the organization’s priorities and risk appetite. More importantly, the AI governance framework must include mechanisms for evaluating and prioritizing AI use cases, ensuring alignment with the institution’s strategic objectives and operational requirements.

2. Model risk management: The experience of financial institutions with financial and risk models provides a foundation upon which AI-specific model risk management practices can be built. However, the introduction of AI technologies, particularly those with autonomous capabilities, requires a reassessment of traditional risk management frameworks. Financial institutions must adopt enhanced risk management strategies that account for the unique characteristics of AI models, including the potential for generative AI technologies to produce novel, sometimes unpredictable outputs.

Strategies such as imposing limitations on data inputs and incorporating human oversight of model outputs are essential for mitigating risks and ensuring the long-term reliability and integrity of AI applications.

3. Centralized standards: To effectively manage the balance between innovation and control surrounding AI, financial institutions must develop and enforce centralized standards. These standards should cover a range of considerations, including ethical use policies, technical development guidelines and governance protocols to oversee AI usage. Establishing centralized oversight ensures that AI initiatives are adopted and implemented in a consistent and controlled manner, facilitating seamless integration into the institution’s operation and IT environment.

For financial executives, the transition toward AI-enabled operations requires careful planning and the establishment of robust foundations in governance, risk management and standardization. By addressing these critical areas, financial institutions can navigate the complexities of AI adoption, ensuring that these technologies contribute positively to operational efficiency, risk mitigation and overall competitive advantage.


Jim Tarantino


Jim Tarantino is a director at RSM.