Skip to Main Content

Artificial Intelligence (AI) Guide

AIRC: AI-Integrated Research Cycle

Artificial Intelligence is revolutionizing (and will continue to revolutionize) the conduct of clinical and basic sciences research. We believe well-equipped users can responsibly implement AI at various stages of the research cycle—in what we refer to as the AIRC (AI-Integrated Research Cycle). The foundations of the AIRC model are these:

  • All Library guidance is superseded by UT systemwide and institutional (VPN or on-campus network) policies, publisher guidelines, and federal and state laws, specifically relating to copyright.
  • Avoid any institutionally-prohibited software (VPN or on-campus network), such as DeepSeek.
  • All AI tools and outputs should be critically evaluated. The predictive, rather than factual or truth-based, nature of AI models means they can always hallucinate, producing false or misleading information.
  • AI-powered tools are research assistants. They are not meant to supersede your expertise, insight, and judgment, whether in an academic, laboratory, or clinical setting. They are meant to enhance, not replace, the traditional research and literature review processes.
  • Library databases are already equipped with AI-enhanced features and continue to make improvements to integrate AI further, thereby enhancing accessibility and furthering research capabilities. External tools, once critically evaluated and utilized ethically, can also assist with research tasks.

AI can be integrated responsibly at each stage in the research cycle:

  1. Create & Plan
  2. Discover
  3. Gather & Evaluate
  4. Write & Publish
  5. Share/Impact

We also provide resources in the following subarea(s) of research:

  1. Clinical

ISAC Approval

All information system, related equipment, and software acquisitions, including AI, must go through the Information Systems Acquisition Committee (ISAC) for approval.

Please use the below links or contact ISAC if you have questions about the process and requirements.

Attribution

Many thanks to the McMaster University Health Sciences Library, which gave us permission to use their research cycle as a template to develop ours.