A Practical Ethics Framework for AI in Health Information Practice: BEST-P

Submitted by: Gabriel Rios

AI is already part of our work, whether we planned for it or not. It shows up in search interfaces, citation tools, writing support, teaching conversations, vendor demos, research workflows, and questions from faculty, students, clinicians, and administrators. Some of these uses are promising. Some are risky. Most are somewhere in between.

That reality was clear at the Medical Library Association (MLA) Annual Meeting in Milwaukee in May. AI came up everywhere: in paper presentations, posters, continuing education sessions, hallway conversations, and in the questions people asked between sessions. The energy was real, but so was the caution. Our community is not simply asking what AI can do. We are asking how these tools should be evaluated and used, and what our responsibilities are when they enter health information environments.

MLA’s new BEST-P Framework gives us a practical way to approach those questions. BEST-P stands for Bias, Expertise, Sustainability, Transparency, and Privacy & Property, and acts as a starting point for assessing how these key considerations may surface in health information contexts. The BEST‑P Framework is not a prescriptive checklist, but rather a flexible, reflective tool that equips health information professionals with a common set of questions to guide collaborative discussions, instructional efforts, vendor assessments, evidence synthesis, and routine decision‑making.

The framework was developed by the MLA AI Imperative Task Force, whose members include Marie Ascher, Dianne Babski, Kristi Holmes, Sarah Jewell, Liz Kellermeyer, Nicki Mehall, Melissa Rethlefsen, Gabriel Rios, and Michelle Rodell. In April 2026, the MLA Board of Directors formally approved the BEST-P framework, recognizing its value as a guiding resource for health information professionals navigating the rapidly evolving AI landscape. The framework itself was coordinated and drafted by Sarah Jewell and Michelle Rodell, with contributions and feedback from the full task force. The group also benefited from the expertise of Mohammad Hosseini, MA, PhD, whose scholarship focuses on the ethical implications of artificial intelligence in research and scholarly publishing, research ethics and integrity, and open science. Together, this collaborative effort grew out of a need many of us have felt across the profession: a clear, practical framework that helps us evaluate and use AI thoughtfully without overstating its potential or overlooking its risks.

MLA’s Role

It is important for MLA to take a leadership role because our members need more than excitement, fear, or scattered advice. We need shared language, practical tools, and places to learn from one another. BEST-P is one part of that work.

MLA’s support for this work also reinforces the important role that medical librarians and health information professionals play in AI efforts, including our early engagement as initiatives develop on campus. We belong in the room when tools are evaluated, policies are written, workflows are redesigned, and users are taught how to engage with new systems. 

Using BEST-P Locally

We hope that MLA members will spend time with the BEST-P resource and ask how it fits their own institutions. BEST-P can support a wide range of activities in libraries: staff discussion and training opportunities, updated vendor review processes, development of student workshops, research consultations, and support for AI conversations with campus and hospital leadership.

The framework will become stronger as we use it and share what we learn. Even though AI will keep changing, our professional expertise and values give us a strong foundation for continued work.

The path forward is not to reject AI outright or to accept it uncritically. Our task is to be careful, informed, collaborative, and willing to keep learning. BEST-P gives us a grounded place to start.