CRIV CONNECTION

Author: Laurel Moran, Assistant Director, Legal Information Management, San Diego Law Library

At the 2025 American Association of Law Libraries (AALL) Annual Meeting & Conference, a timely and insightful program addressed the responsibilities of librarians in evaluating and licensing generative artificial intelligence (AI) technologies. The panel featured Jennifer Wondracek (Director of the Capital University Law School Law Library); Kacie Weiper (Digital Services Librarian at Perkins Coie LLP); and Karen Whitney (Library Resources Manager at Perkins Coie LLP).

A recording of the program is available on AALL eLearning.

The presentation opened with a short definition of generative artificial intelligence. These products help their users create and transform content and support interactive work and analysis. Licensing generative AI products is of heightened concern as the vendor license terms often trigger issues of data ownership, data privacy, and data security. The speakers reviewed six major factors for librarians to consider when reviewing licensing agreements for Gen AI products.

The first of these considerations is data usage and ownership. The speakers advised librarians to view each licensing agreement and identify who retains ownership of inputs (AI prompts or documents uploaded). It is essential to ask vendors who retains the ownership of the inputs, how vendors use the inputs, and how the client input is used in training the generative AI product. Make sure you know if you can opt-out of certain uses of your data before entering a licensing agreement.

Next, the panel covered data privacy and security. The speakers encouraged librarians to prepare tough questions for vendors on this issue. Librarians should ask vendors where user-submitted data is stored and whether it is shared with third parties like Open AI. Know that if your user’s data ends up in Open AI, it may be subject to electronic discovery later. Librarians should also understand the security mandates of their firms, government employers, or schools, such as government procurement rules and Federal Education Record Protection Act (FERPA) requirements.

Accessibility was the third major topic. Librarians must ensure that any generative AI product complies with their organization’s accessibility standards. This includes knowing what documentation is required and communicating those needs to vendors early in the negotiation process to avoid delays.

A fourth consideration is liability and risk. Librarians need to understand how a product will be used and the professional standards of their library users. If a librarian can spot red flags or risks in advance when working with their vendors, those risks may be covered in training or in the setup of the product. The speakers emphasize that strong vendor-librarian relationships are particularly important to mitigate liability and ensure the smooth implementation of a product.

The speakers mentioned that they proactively manage updates to generative AI products and always consider how to communicate with library stakeholders. These products are updated so often with new functionality and features that librarians can get overwhelmed with how often to communicate changes. Significant product updates and new features may also trigger a new security review, so librarians are advised to pay attention to the sea of update notices they receive with these products. Effective communication is also key to informing library stakeholders about products, product updates, and functionality. Build relationships in advance with your IT and security teams to ensure that product updates do not disrupt workflows.

Finally, the speakers emphasized the importance of post-licensing implementation. Once a license is signed, librarians should communicate updates to users, require user training before granting access, and use pop-up reminders (on product sign on) for institutional AI use policies.

In conclusion, the program provided a practical framework for librarians to evaluate and implement generative AI products. By asking the right questions and building strong relationships with vendors and stakeholders, librarians can play a pivotal role in guiding their institutions through the complexities of AI adoption.