Artificial intelligence (AI) is no longer a futuristic concept for borrowers. It is already embedded in products, operations, data management, and decision-making across industries. For lenders, this means that AI is not simply a technology issue. It can directly affect enterprise value, regulatory exposure, intellectual property rights, and ultimately, repayment risk.
While traditional credit agreement provisions addressing data security, intellectual property, litigation, and material adverse effect remain relevant, a borrower whose business materially involves AI or large language models warrants a more deliberate and thorough review through a credit risk lens.
Diligence Considerations
When evaluating and documenting a credit facility for a company that develops, deploys, licenses, or materially relies on AI, lenders should tailor their approach to diligence and documentation to capture AI concerns. From a diligence perspective, they should evaluate how AI drives revenue, how models are trained and maintained, and how dependent the business is on third-party data, infrastructure, or tools. AI may be central to revenue generation, integral to the company’s data acquisition and monetization practices, or a significant source of regulatory and litigation exposure. In some business models, it may also affect the value, transferability, or enforceability of core intellectual property that underpins the lender’s collateral package. These issues may not be adequately captured by standard diligence checklists and underwriting practices.
Documentation Considerations
From a documentation perspective, lenders may consider whether the compliance with law covenant adequately addresses applicable AI regulation and regulatory guidance, particularly where the borrower operates in highly regulated sectors or across multiple jurisdictions. Representations may warrant refinement to address the borrower’s rights to use customer data to train AI systems, the legality of the borrower’s data sourcing practices, the absence of known model misuse, and the infringement or misappropriation of third-party intellectual property by the borrower’s AI systems and related outputs. Where AI forms a material part of the business, lenders may also evaluate whether additional disclosure or reporting is appropriate with respect to regulatory inquiries, material model failures, significant data incidents, or AI-related litigation.
Governance and Covenant Protections
In extending credit to borrowers whose products or services depend heavily on AI, lenders may consider including covenants that require the borrower to maintain reasonable AI governance frameworks, internal controls, and risk management policies. These provisions can serve not only as compliance tools but also as early warning mechanisms where weaknesses in oversight could translate into operational disruption or reputational harm.
Market Practice and Forward Outlook
AI-focused drafting is unlikely to become a uniform market standard in the immediate future. The appropriate approach will depend on the borrower’s specific business model and risk profile. However, where AI is central to enterprise value, tailored credit agreement provisions can enhance transparency, strengthen risk allocation, and better align documentation with the realities of the borrower’s operations. Thoughtful drafting at the outset can reduce uncertainty, support underwriting assumptions, and position lenders to respond more effectively as regulation and market practice continue to evolve.
[View source.]
