Artificial Intelligence

Position

• Banks, as early adopters, have effectively integrated AI and machine learning into their operations, prioritizing safety and soundness.
• ICBA holds that banking regulators currently possess the necessary tools to oversee AI usage in community banks.
• ICBA believes that regulations that currently apply to the banking sector should be extended to other industries to provide similar protections to consumer data and privacy.
• ICBA opposes laws, regulations or guidance that could lead to the commoditization of lending and other services and products offered be community banks.
• ICBA opposes excessively prescriptive regulations from the CFPB that could potentially impede the use of AI in developing more equitable lending models.

Background

Banks have a long history of employing machine learning and artificial intelligence (AI) to meet regulatory requirements and enhance customer service. AI has been used to detect and prevent fraud and ensure compliance with the Bank Secrecy Act, and chatbots have long been used in call centers.

Following the emergence of ChatGPT and similar generative AI tools, there has been a notable surge in interest regarding AI applications across various industries, including banking. Though Congress and federal agencies are actively exploring AI-related policies, substantive modifications or policy changes have not been officially adopted. ICBA remains engaged with policymakers to ensure that their focus does not result in laws or regulations that would hinder community bank adoption of AI for internal or external applications.

Existing regulations and supervisory guidance on model risk management, fair lending, and data privacy/security offer ample flexibility to regulate AI use by financial institutions. ICBA believes that AI-specific regulations are unnecessary at present and that federal regulation would only be appropriate in response to significant technological changes or to preempt conflicting state regulations. Additionally, ICBA emphasizes that AI, though useful in automating community bank compliance functions and loan underwriting, cannot replace the personal relationships and local knowledge that are integral to the community banking model.