Institutional AI Policy Draft

University of Tennessee Health Science Center
Prepared by the Institutional AI Strategy Task Force

1. Purpose
   This policy provides the framework for the responsible and secure application of artificial intelligence (AI) technologies in support of the core missions of the University of Tennessee Health Science Center (UTHSC). The purpose is to guide faculty, staff, students, and partners in upholding our commitment to advancing healthcare and the health of Tennesseans through innovative technology while maintaining ethical standards.
2. Definitions
   Artificial Intelligence – A machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments and that can use machine and human-based inputs to perceive real and virtual environments, abstract such perceptions into models through analysis in an automated manner, and use model inference to formulate options for information or action. (State of Tennessee Public Chapter No. 550 Senate Bill No. 1630)
   
3. Scope
   The policy applies to all AI and AI-driven products and services deployed for university-affiliated activities, including academic tools utilized for both teaching and learning, clinical/patient care systems, administrative product and services, research initiatives, and community engagement platforms. It also covers multiple forms of AI, including generative AI as well as AI that is integrated into other processes or products. Finally, this policy encompasses ethical considerations, governance and oversight, risk and security, compliance, and technical implementation. This policy seeks to balance the benefits of AI with responsible use in accordance with applicable laws. In instances where this policy conflicts with Board of Trustees-approved policies, or federal or state law, those policies supersede this policy.
4. Standards of Use
   AI use at UTHSC will adhere to the following standards:
   1. Ethical Principles: Adherence to these ideals is paramount in the deployment and development of AI technologies.
      1. Respect for individuals: Uphold and protect individual dignity and rights by ensuring our use of AI products and services treats all users with respect, recognizing their inherent value as human beings.
      2. Beneficence: Actively ensure that our AI products and services contribute to positive health, research, and educational outcomes, enhancing well-being while avoiding harm.
      3. Justice: Commit to providing equitable access to the benefits of AI products and services across different demographics, ensuring that no group is disadvantaged by technological advancements.
      4. Transparency: Strive for transparency in our use of AI products and services, making them understandable to non-experts and appropriate, thus building trust and accountability.
   2. Data Governance: Data governance is critical for maintaining user trust, regulatory compliance, and ethical use of AI products and services.
      1. Each user has a responsibility to guard personal privacy through stringent adherence to laws like HIPAA for healthcare data and FERPA for educational records.
      2. University standards for data accuracy, consistency, and reliability must be upheld during all stages of collection and processing.
      3. Comprehensive protocols must be established that outline conditions under which data may be shared with external entities while safeguarding sensitive information.
   3. Compliance: Compliance ensures the University operates within legal frameworks.
      1. The university will comply with current laws and regulations governing AI use, recognizing this as a foundation of ethical practice.
      2. Regular policy review processes will be instituted in anticipation of changes in legal standards, ensuring our practices remain current.
   4. Transparency and Consent: Transparent operations build trust; consent ensures autonomy.
      1. Use of AI will be disclosed in significant decision-making processes involving personal records or healthcare decisions.
      2. When protected or sensitive data is being used, explicit consent from individuals will be obtained before utilizing their data or materials in AI products and services whenever it is required by law or best practice standards.
   5. Accountability and Oversight: Accountability structures are critical for responsible stewardship of AI.
      1. The university will implement a governance structure to monitor all facets of AI and AI-related initiatives within the university's ecosystem.
      2. All AI users must understand their responsibility regarding use of AI products and services.
      3. Specific domains or units of the university may need to compose sub-policies to govern AI use. For instance, academic leaders may choose to define what is acceptable use or citation of AI-generated content in student work.
   6. Security and Risk Management: Security measures prevent misuse; risk assessments anticipate challenges.
      1. Appropriate cybersecurity measures will be implemented to protect against unique vulnerabilities presented by sophisticated AI systems handling sensitive data.
      2. The university will conduct thorough risk assessments regularly on both existing and proposed AI systems to identify potential threats early on.
   7. Academic Accountability: Academic accountability assures the integrity of teaching and learning.
      1. Integration: AI technologies should be integrated into academic courses where they can demonstrably improve instructional quality, student engagement, or learning outcomes.
      2. Transparency: The use of AI must be transparent; students and faculty should be informed about how these technologies function and their role within the course. Faculty should clearly state course-specific AI policies in the course syllabus.
      3. Privacy: All AI products and services must comply with university privacy policies as well as federal and state regulations concerning data protection.
      4. Bias Mitigation: Efforts shall be made to address and minimize the bias inherent in AI products and services that could affect student assessment or access to learning opportunities.
      5. Accessibility: AI resources must adhere to accessibility standards so that all students have equal opportunity to access and benefit from their use.
      6. Accountability: Faculty remain ultimately responsible for course content and quality assurance; reliance on AI does not diminish faculty oversight or accountability.
      7. Compliance: Students are responsible for adhering to academic integrity policies and policies related to the use of AI as specified by their professor, the department, the college, and UTHSC.
   8. Research Integrity and Collaboration: Integrity in research fosters trust; collaboration extends boundaries.
      1. The university will align research involving AI with recognized scientific integrity principles, promoting accuracy, objectivity, transparency, and carefulness of work.
      2. Research collaborations will honor this policy’s tenets, ensuring partnerships reflect a commitment to ethical practice.
   9. Training and Awareness: All AI users must understand the principles, practices, and implications associated with AI technologies.
      1. Training resources that reflect the principles outlined in this policy will be available for both AI users and developers.
      2. Widespread understanding of ethical considerations related to AI will be promoted throughout the university community via seminars, workshops, and informational campaigns.
   10. Continuous Improvement: Adaptation is essential; feedback informs evolution.
       1. The university will engage robust feedback mechanisms between users, developers, administrators, and other stakeholders to refine and improve existing AI systems.
       2. This policy will be periodically reviewed and updated based on new insights, technological advancements, and regulation changes.
   11. Technology Implementation: University leaders, stakeholders across the university, and the IT department will collaborate to identify, prioritize, fund, and implement AI technologies. All AI decisions will support the university’s strategic plan.
5. Review
   This policy will be reviewed annually by the designated governance structure to remain relevant, effective, and aligned with the UTHSC mission.
6. Inappropriate Use
   Inappropriate use of AI may result in disciplinary action consistent with the University’s code of conduct, acceptable use, honor code, and other applicable policies and standards.