Responsible Use of Generative Artificial Intelligence
Introduction
Generative artificial intelligence (GenAI) has the potential to transform many aspects of learning and working, including the University experience for students and for employees. This policy is intended to guide the responsible use, development, and management of GenAI at the University of Regina.
The University of Regina recognizes four dimensions of integrating GenAI into university operations: Pedagogical, Research, Governance and Operational. The Pedagogical dimension considers how GenAI may be used to improve teaching and learning outcomes. The Research dimension considers GenAI as a focus of research itself, and as a tool that may be applied to existing research practices. The Governance dimension provides guidance related to privacy, security, compliance and ethics. The Operational dimension addresses matters concerning infrastructure and employee training.
The Pedagogical dimension is within the oversight of existing academic governance structures and processes (e.g. Executive of Council, Senate, and committees of the Office of the Provost).
The Research dimension is within the oversight of the Office of the Vice-President Research. Employing AI in primary research is governed by all the same policies and regulations that govern non-AI-assisted research; however, Canadian funders (NSERC, SSHRC, CIHR, and the CFI) have also issued “Guidance on the use Artificial Intelligence in the development and review of research grant proposals”.
This policy does not apply to research projects where relevant approvals and ethics clearances are covered by research policies, although any such uses should be consistent with the intent of this policy.
This policy is focused primarily on the remaining Governance and Operations dimensions relating to GenAI. This policy applies to the development, approval, use and management of GenAI software, systems, or platforms that may be used by University of Regina employees, with a particular focus on providing guidance for non-academic employees and purposes.
Definitions
- AI – artificial intelligence (AI) refers to information technology that performs tasks that would ordinarily require biological brainpower to accomplish, such as making sense of spoken language, learning behaviours or solving problems.
- Generative AI (GenAI) – is understood to mean a particular type of AI that produces content such as text, audio, computer code, videos or images based upon data used to train the underlying model. Current examples include ChatGPT, CoPilot, and DALL-E.
Policy
All University use of GenAI must be ethical, reliable, transparent, secure and compliant with applicable laws and regulations.
To maintain public trust and ensure the responsible use of GenAI technology and tools, the University mirrors Government of Canada’s “FASTER” guiding principles summarized as follows:
- Fair
- We endeavor to ensure that GenAI content does not include or amplify biases and that it complies with our commitments to uphold human rights, accessibility, and procedural and substantive fairness
- We engage with affected stakeholders before deployment
- Accountable
- We take responsibility for the content generated by GenAI tools and the impacts of their use
- We establish monitoring and oversight mechanisms
- This includes making sure generated content is accurate, legal, ethical, and compliant with the terms of use
- Secure
- We ensure that the infrastructure and tools are appropriate for the security classification of the data provided or the content produced
- We ensure privacy, intellectual property, and personal information are protected
- We assess and manage cybersecurity risks and robustness when deploying GenAI systems
- Transparent
- We identify content that has been produced using GenAI
- We notify users that they are interacting with an GenAI tool
- We are able to provide information on appropriate use, training data and models
- We document decisions and are able to provide explanations if GenAI tools are used to support decision-making
- Educated
- We learn about the strengths, limitations and responsible use of GenAI tools
- We learn how to create effective prompts and to identify potential weaknesses in the outputs
- Relevant
- We make sure the use of GenAI tools and systems supports organizational needs and contributes to better outcomes
- We identify appropriate tools for the task and recognize that GenAI is not the best solution in every situation
The University is committed to implementing best practices — such as maintaining good data hygiene, using high-quality data sets, empowering users with control over their data, following established governance frameworks, and monitoring implementation. Through this, the University will effectively integrate GenAI into operations while safeguarding data privacy and security, and maintaining compliance with regulatory requirements and ethical standards.
The University acknowledges that:
- there are legitimate concerns about the use and reliability of GenAI. The University will seek to identify the balance between seeking opportunities for GenAI integration, while managing and mitigating risks to the university and its community at all times.
- algorithmic bias may result in erroneous or unjustified differential treatment which could have unintended or serious consequences for individuals and/or for human rights.
- the use of GenAI will be increasingly regulated and, as such, this policy and any associated procedures must be reviewed regularly and maintained to ensure compliance with current and emerging regulatory standards.
The University will continuously manage and evaluate the use of GenAI to effectively balance risks and opportunities. The University also recognizes that integration of GenAI will require a level of experimentation as well as ongoing enhancement of individual-level competencies and skills and will encourage use of GenAI technology as a supplementary tool, not as a replacement for human interaction.
Data used to develop algorithms or GenAI systems, and any data generated, shared, managed and/or recorded as part of an GenAI system’s operation or algorithm, will be considered institutional data and must be managed in alignment with GOV-060-005 Freedom of Information and Protection of Privacy policy and the University’s data handling standards. This requirement will apply to the full project lifecycle and to the lifetime of the data (whichever is longer).
Continuous learning and improvement will inform future use of GenAI technology and tools across the University.
Roles and Responsibilities
President, Vice-Presidents, and Chief Governance Officer
- Establishing and maintaining this policy
- Allocating resources to adequately support GenAI-related projects
- Providing oversight and approval for University-wide GenAI projects
Associate Vice-President (Information Services)
- Ensuring approved GenAI-related tools and systems are assigned to individuals with relevant expertise with the technology
- Maintaining a central record of approved University-wide GenAI tools and systems that have been reviewed by the Technology Risk Management Committee (TRMC)
University Employees (Faculty and Staff)
- Integrating ethical considerations into all GenAI-related activity
- Demonstrating and fostering transparency in use of GenAI
- Complying with this policy
- Establishing processes to allow affected users to ask questions and seek reviews of GenAI-based decisions
- Providing human review and intervention for GenAI-based activities and decisions
- Ensuring any use of GenAI in operations is properly documented and referenced
Consequences for Noncompliance
Failure to comply with this policy may have consequences for the individual and for the University, including but not limited to: disciplinary action; legal action; imposition of fines/financial penalties; and potential loss of reputation.
Processes
Faculty- and/or Unit-level leaders who are considering a GenAI use case should maintain an open dialogue with relevant University experts and stakeholders as part of GenAI project approval, implementation and ongoing management. This dialogue may include the Offices of Information Services, University Governance, Financial Services, Institutional Research, and data stewards and information system stewards as appropriate.
Where a GenAI system is being proposed for either development or procurement, the normal project approval and procurement processes should be aligned with the GenAI project approval process wherever possible (e.g. privacy impact assessment) to avoid duplication of effort.
All GenAI systems exist on a spectrum of risk, ranging from low-risk (not automated, rarely operational, does not contain personal or sensitive data, within substantial controls, and does not have direct impacts on individuals) to high-risk (highly autonomous, normally operational GenAI systems, containing personal or sensitive data, with minimal controls, and could immediately or substantively impact individual and institutional safety and wellbeing).
In alignment with the University’s commitment to risk management, any employee who is considering a GenAI project or system must identify, assess and analyze risks and opportunities, including alignment or compliance with current University policies and/or legal and regulatory requirements, including privacy and security.
Any University-wide GenAI project, tool, system or intended use of enterprise-level data requires approval through established IT project approval processes (see OPS-080-030 Information Technology Initiatives).
Related Information
- Guide on the use of generative AI (Government of Canada)
- Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems (Government of Canada)
- University of Regina AI Guidelines for faculty and instructors
- University of Regina AI Guidelines for meeting minutes
- GOV-060-005 Freedom of Information and Protection of Privacy
- OPS-080-030 Information Technology Initiatives
- GOV-022-005 Code of Conduct
- GOV-050-010 Use of Copyrighted Materials