The Columbia Shuswap Regional District (CSRD) is setting guidelines for use of artificial intelligence.
For September's board meeting, CSRD staff will be bringing forward updated policies and procedures that will address the adoption of artificial intelligence (AI) applications, in particular the use of Microsoft CoPilot. Prior to the board meeting, however, staff will be presenting a draft policy for AI usage at the Aug. 7 Committee of the Whole meeting.
"The purpose of this policy is to establish guidelines for the acceptable use of Artificial Intelligence (AI) technologies within the Columbia Shuswap Regional District (CSRD)," reads a draft policy titled Acceptable Use of Artificial Intelligence. "Through the responsible use of AI, the CSRD can drive innovation, increase operational efficiencies, while protecting privacy, managing risk, and promoting accountability, safety, and equity."
The policy includes guiding principles around the use of AI, data privacy and security, "transparency and explainability, bias and fairness, human oversight and compliance with laws and regulations.
Regarding transparency and explainability the document states AI systems, "especially those affecting critical business decisions, should be transparent and understandable," and that the "decision-making process of AI systems must be explainable, and employees, elected officials, or stakeholders should be able to understand the rationale behind AI-driven decisions."
Human oversight is expected to be maintained over all AI systems, with humans always having the option to "intervene in AI-driven decisions when necessary."
The policy calls for the creation of an AI ethics committee to oversee AI projects and ensure compliance with the policy.
A second draft policy, titled Artificial Intelligence Usage, delves into more specific do's and don'ts for CSRD employees and elected officials. Regarding the protection of sensitive information, the policy states confidential information may be put into "Generative AI (GenAI) tools available to them as part of the BC Public Service鈥檚 agreement with Microsoft, with M365 Copilot Chat being the recommended product to be used. However, confidential information cannot be put into publicly available AI tools like ChatGPT.
The policy also requires employees and elected officials who use AI to "review and edit the output text to ensure the language is inclusive, respectful, and aligns with government style guides and other writing standards." The accuracy of information provided by AI must also be verified.
"The use of M365 Copilot Chat is to supplement your work, do not use the program to replace human judgment and creativity," reads the policy. "Apply critical thinking to the responses provided by M365 Copilot Chat and make decisions based on a combination of AI input and human expertise."
According to staff, AI can benefit the regional district's ability to "analyze large datasets for things like planning and environmental monitoring," automate routine administrative tasks and help reduce operational costs through automation by "enhancing productivity without expanding workforce" and "supporting collaboration across departments and municipalities."
