Use of Generative Artificial Intelligence (Policy 94)

1. Governors State University (GSU) is committed to utilizing current and future technology, including Artificial Intelligence (AI), in the pursuit of its mission of offering an exceptional and accessible education that prepares students with the knowledge, skills and confidence to succeed in a global society. The use of generative AI has implications for the university’s community standards pertaining to academic pursuits, including fulfilling academic requirements and scholarship/creative endeavors. This policy is meant to describe how decisions about the use of generative AI should be made, as well as present guidelines on the appropriate and inappropriate uses of generative AI. Principles of academic freedom and responsibility described in GSU’s Board of Trustees Governing Policies are primary considerations for the standards of the use of generative AI at Governors State University. Consistent with these principles, decisions about the use of generative AI are made by the individual faculty members responsible for the pursuit of excellence in instruction and scholarship/creative endeavors. However, common considerations of integrity and fairness also exist across academic disciplines that are relevant to the use of generative AI. Accordingly, this policy also provides guidelines for promoting academic honesty and managing risks associated with the use of generative AI.

Definitions

AI is a rapidly developing technology involving many terms that may be novel to those with limited experience working with it. As the implications of generative AI for higher education broaden, a clear understanding of terms associated with AI is important for accurate consideration of standards for its use. Listed below are key definitions for understanding generative AI.

1. “Artificial Intelligence” (AI) is a form of technology that allows machines to mimic a range of sophisticated human skills. 
2. “Machine Learning” (ML) is a branch of AI that enables computers to acquire knowledge and improve from experience without the need for specific programming. 
3. “Generative AI” is a subset of ML technologies that can use the data it has been trained on to create new content, such as writing, images, audio, and video. Commonly used applications of generative AI include ChatGPT, Microsoft Copilot, and Google Gemini. 
4. “Artificial General Intelligence” (AGI) is a conceptualization of technology possessing autonomous sentience that most AI experts agree has yet to be achieved. AGI is distinct from generative AI and these terms should not be confused or conflated with one another.

1. Policy Scope

   1. Description 

      This policy is specifically designed for decisions of using generative AI in academic pursuits, such as fulfilling academic requirements and scholarship/creative endeavors. This policy is intended to be consistent with the Board of Trustees Governing Policies and their statements on academic freedom and responsibility which supports the notion that the discretion to use generative AI appropriately rests with the faculty members, who are dedicated to fostering high standards in education and research or creative activities. This approach is in harmony with their role in upholding the highest standards of teaching and advancing scholarship or creative pursuits. This policy is also intended to be consistent with other GovState policies pertaining to academic pursuits, including but not limited to policies on Academic Honesty (Policy 24) and Research and Scholarship (Policy 55). 

      There are many non-academic pursuits for which generative AI may be used, including personal communications, internal business communications, and works not intended for fulfilling academic requirements or academic publication or presentation. Such non-academic pursuits are beyond the scope of this policy.

   2. Sanctions

      Any sanctions for inappropriate use of generative AI will be determined by the application of existing policies associated with specific aspects of misconduct in academic pursuits, including Student Conduct (Policy 4), Academic Honesty (Policy 24), Anti-Discrimination, Harassment, and Retaliation Policy (Policy 52), Research and Scholarship (Policy 55), Fair Use of Copyrighted Works for Education and Research (Policy 62), and Responding to Allegations of Research Misconduct (Policy 71).

2. Faculty Academic Freedom and Responsibility and the Use of Generative AI Principles of academic freedom and responsibility described in GSU’s Board of Trustees Governing Policies recognize faculty as experts in their disciplines with an obligation for upholding the integrity and quality of their academic programs. Faculty are responsible for staying current in their disciplines and as the use of generative AI proliferates within various academic domains faculty have the obligation to guide its use in educational pursuits in accordance with the standards within their respective disciplines. Technological advances, like generative AI, need to be accessible to faculty and students so that they may apply them to their respective disciplines. Moreover, all members of the academic community have an obligation to promote and protect intellectual honesty and freedom of inquiry while fostering an environment for learning that supports individual rights and maintains the dignity of others. As generative AI is integrated into processes of inquiry and creative expression within various disciplines, considerations of its appropriate and inappropriate use must be guided by disciplinary expertise and the pursuit of excellence and integrity for instruction and scholarship/creative endeavors.

3. Academic Honesty and the Use of Generative AI

   Ethical standards for maintaining the integrity of academic pursuits are described in Policy 24: Academic Honesty which “pertains to all methods of fulfilling academic requirements at Governors State University”, including the use of generative AI (see also Policy 4: Student Conduct). In accordance with these standards, students may use generative AI tools with instructor’s permission. It should not be stated or implied that the output of generative AI is a person’s own work. Direct quotation of material from generative AI should be enclosed in quotation marks or otherwise set off, and the generative AI source should be acknowledged. Paraphrasing of material from generative AI should also acknowledge the generative AI source used. The content produced by generative AI applications may not adequately cite its source material and therefore may not adequately fulfill academic requirements for using the ideas and works of others. Those using generative AI should seek to identify primary sources used in the content it provides and confirm the accuracy of these citations.

4. Managing Risks Associated with the Use of Generative AI

   1. Ensuring the Veracity of Generative AI Output

      While generative AI is often effective at generating content, it also is prone to creating inaccurate outputs, a process commonly described as “hallucinating”. For instance, generative AI can fabricate details in its outputs. Mitigating risks of inaccuracies is therefore a key consideration for the use of generative AI. Prudent selection of generative AI products is one method of ensuring the veracity of generative AI output. Transparency about key components of the AI process enables the prioritization of products with features that minimize the risk of “hallucinating”, including:  training the AI on high-quality, diverse, and accurate data sources and using algorithms designed to fact-check output. Furthermore, users of generative AI should be aware of the limitations of the specific tool being used and critically evaluate its output by cross-verifying it with independent sources that are trusted and authoritative.

   2. Data Privacy Risks

      The use of generative AI can pose a number of risks to the privacy rights of students and due consideration should be given to using generative AI in a manner that protects student privacy. Using generative AI to tailor individual student learning experiences might capture data on student academic performance or other data for which privacy is protected (e.g., status of learning disabilities or mental health). Thus, AI products used in academic pursuits should have transparent disclosures about data use that ensure robust data protection measures that comply with GSU’s standards for data privacy (see Policy 61: Privacy, Legal Notices, and Security Notification), as well as applicable privacy laws and regulations and best practices for using personal data.

   3. Data Bias Risks

      The process of training generative AI systems can produce biased results when they involve historically biased or unrepresentative datasets. Mitigating the risks of such bias should be a primary consideration for the application of generative AI in aspects of education that include but are not limited to (see also Policy 52: Anti-Discrimination, Harassment, and Retaliation Policy): 

      1. Application and admission to the University; 
      2. Treatment in the classroom; 
      3. Academic activities external to the classroom; 

      The development, deployment, and evaluation processes for generative AI systems within the scope of this policy should seek feedback from a broad range of stakeholders. Identifying generative AI products that are transparent and make the criteria and algorithms used by the AI clear to users can aid in mitigating risks of data bias. Users of generative AI should seek products that take measures to mitigate bias, such as (1) training the generative AI on data that is representative of various demographic groups and geographic locations, (2) using algorithms designed to detect and mitigate bias in AI systems, (3) implementing regular, third-party audits that help to identify and correct biases, and (4) employing user feedback mechanisms to identify and redress potential bias.

   4. Copyrighted Works and Generative AI

      By using generative AI products that clearly disclose their data sources and algorithmic processes, users can more easily identify and mitigate risks associated with copyright infringement. Use of generative AI should be in accordance with the University policy on the “Fair Use of Copyrighted Works for Education and Research” (Policy 62) and intellectual property laws of the United States. Use of generative AI can produce content that is derivative of copyrighted material and potentially infringes on the original copyright. Standards of Fair Use allow students, educators, and researchers to use copyrighted materials for learning and knowledge advancement in ways that might otherwise infringe on copyright laws. Users of generative AI should be mindful of the evolution of legal precedents associated with the use of generative AI, including those associated with Fair Use.

   5. Research Integrity and the Use of Generative AI

      Ethical standards for maintaining the integrity of research and scholarship are outlined in Policy 55: Research and Scholarship. Key components of Policy 55 are assurance of open scholarly exchange and academic freedom which must be upheld and extended to include the use of generative AI. While academic freedom is a primary consideration for the use of generative AI, it should not be misconstrued as justifying academic misconduct (see also Policy 71: Responding to Allegation of Research Misconduct). Aspects of academic misconduct pertinent to the use of generative AI include, but are not limited to: 

      1. Fabrication or falsification of data with generative AI to deliberately report false results; 
      2. Confidential data, including non-public research data, may not be submitted to generative AI tools when such submission could expose sensitive, private, or proprietary information; 
      3. Unacknowledged appropriation of the work of others with generative AI, including plagiarism. 

      Researchers assure appropriate use of generative AI primarily by self-regulation and by individual adherence to ethical codes and standards within their disciplines of expertise. Professional associations may have ethical codes and guidelines for the use of generative AI in the conduct of research; university personnel are expected to comply with such standards.