Guidance on Artificial Intelligence use

IST supports the innovative use of technologies to solve problems and optimize processes which support the mission and business operations of the University of Waterloo. Modern technologies which fall under the umbrella of Artificial Intelligence (AI) are becoming more accessible to a wider audience and due to the nature in which they work, require extra caution to be used responsibly.

What is AI?

Artificial Intelligence (AI) refers to any system which operates or performs tasks that are normally performed by human intelligence. Under the umbrella of AI sit many fields, but the topic most often talked about is Generative AI (like ChatGPT or Microsoft Copilot).

Systems like ChatGPT or Microsoft Copilot are developed using modern AI technologies, and then made useful to the public by being trained on massive amounts of data. Once trained, they can be used by anyone through something like a chat interface, and respond with anything from text, images, and audio to video.

General policy

While the University works on developing its policy on Artificial Intelligence, IST urges proceeding with caution and due diligence while using University data to train, prompt, or interact with AI systems that are not yet licensed or vetted.

All existing University policies still apply to the usage of AI, in the same way they do to all information systems. Ensure an understanding of the following, policies, for example:

• https://uwaterloo.ca/secretariat/policies-procedures-guidelines/policies/policy-46-information-management
• https://uwaterloo.ca/secretariat/policies-procedures-guidelines/policies/policy-73-intellectual-property-rights
• https://uwaterloo.ca/privacy/information-risk-assessment

Responsible use

If using or developing an AI system, ensure it is done responsibly. Consider aligning with the Government of Canada's recommended approach for the responsible use of AI:

• Fair: ensure that content from these tools does not include or amplify biases and that it complies with human rights, accessibility, and procedural and substantive fairness obligations
• Accountable: take responsibility for the content generated by these tools. This includes making sure it is factual, legal, ethical, and compliant with the terms of use
• Secure: ensure that the infrastructure and tools are appropriate for the security classification of the information and that privacy and personal information are protected
• Transparent: identify content that has been produced using generative AI; notify users that they are interacting with an AI tool; document decisions and be able to provide explanations if tools are used to support decision-making
• Educated: learn about the strengths, limitations and responsible use of the tools; learn how to create effective prompts and to identify potential weaknesses in the outputs
• Relevant: make sure the use of generative AI tools supports user and organizational needs and contributes to improved outcomes for Canadians; identify appropriate tools for the task; AI tools aren’t the best choice in every situation

Or with the Government of Ontario's principles for ethical use of AI:

• Transparent and Explainable: Data-enhanced technologies, including AI, should be used transparently and responsibly. This involves disclosing how and why these technologies are employed. When automated decisions are made, meaningful explanations must be provided to those affected. These explanations should include relevant information about the decision, its process, and its consequences. Transparency fosters trust and encourages dialogue between technology users and those impacted by it.
• Good and Fair: Throughout their life cycle, data-enhanced technologies should respect the rule of law, human rights, civil liberties, and democratic values. This ensures that their design and operation align with ethical considerations and values.
• Safe: Safety is paramount. Data-enhanced technologies must be developed and operated in a manner that prioritizes safety, minimizing risks to individuals and society. This includes safeguarding against unintended consequences and potential harm.
• Accountable and Responsible: Accountability is crucial. Those responsible for deploying data-enhanced technologies should be accountable for their impact. This involves monitoring and evaluating their effects, addressing any adverse outcomes, and ensuring responsible use.
• Human-Centric: Data-enhanced technologies should prioritize human well-being. Their design and implementation should enhance human experiences, promote inclusivity, and consider the diverse needs of users.
• Sensible and Appropriate: Balancing innovation with ethical considerations is essential. Ontario’s principles harmonize with existing best practices and frameworks, aiming to create clarity rather than barriers for innovation. Feedback from stakeholders is encouraged to refine and adapt these principles.

Development

If developing an AI system from the ground up, the following additional checks and balances should be incorporated into the development and deployment workflow:

• Define a clear system purpose
• Design a system which strives for interpretability
• Audit against the system's intended purpose pre and post deployment
• Ensure protocols are in place to understand the quality and structure of the test data used during audits
• Ensure University software best practices are being followed
• Ensure all University Policies are being followed, and where unsure, seek guidance

What tools are available?

The University does not currently provide licenses for any enterprise Generative AI tools.

This list will be kept up to date as new tools arise.