NWT government has no plan to develop AI policy

The NWT government says it is not currently planning to develop a standalone policy on the use of artificial intelligence in the public service.

The NWT government released a high-level guideline on the use of generative AI in May 2025.

Territorial officials say the guideline and existing information management policies are adequate, but some experts say more is needed to address the potential risks of using AI.

Generative AI tools – such as ChatGPT, Dall-E and Microsoft Copilot – can create original content based on prompts including text, images, music and code.

Among its recommendations, the guideline states the territorial government should:

• establish clear rules and responsibilities for using and managing generative AI programs;
• put safeguards in place to protect data and generative AI outcomes and manage risks;
• provide clear information on why, how and when generative AI is used; and
• monitor generative AI programs.

The guideline includes links to the federal government’s guide on generative AI as well as NWT policies on privacy and handling information.

The Department of Finance said GNWT employees have access to an “AI Hub” that includes the guideline, general generative AI training, and training on Microsoft Copilot.

Caroline Wawzonek, the territorial minister of finance, told Cabin Radio the NWT government is “not going to add more to the existing policies that are there.”

She said the public “should have a pretty high level of confidence” in the government’s cybersecurity and its management and protection of information.

“This is a major area of onboard training and consistent training and publications,” she said.

The version of Microsoft Copilot available to GNWT employees is secure, Wawzonek added, so information does not leave the government’s server system. She said the GNWT also expects employees who use generative AI to review the output.

While the NWT government has not conducted its own assessment of the privacy impacts of AI tools, the finance department said it has “leveraged the privacy impact assessments conducted by other jurisdictions and is conducting a legal review of terms of use of common AI tools including Microsoft Copilot.”

Risks and benefits

AI has become ubiquitous in many people’s daily lives, from Google search results to business advertisements and posts on social media.

It is also becoming more common in the workplace, including the public service.

This month, the NWT government announced a pilot project involving an AI-powered medical note-taking tool intended to help reduce the amount of time healthcare workers spend on paperwork.

The Department of Justice told Cabin Radio it launched its own internal pilot more than a year ago involving the use of AI to quickly produce rough transcripts from court proceedings. The department said only court services employees and judges have access to those rough transcripts.

While AI tools offer benefits such as streamlining or automating tasks, there can also be drawbacks including privacy risks and negative impacts on critical thinking and problem-solving skills, as well as the environment. AI can also perpetuate biases and provide inaccurate information, known as AI hallucinations.

The NWT’s wildfire agency, for example, publicly criticized a Facebook page for using an inaccurate AI-generated image on a post about a wildfire burning outside Fort Providence last year. NWT Fire called the image “sensationalized slop” and said those involved “should be ashamed of themselves.”

The province of Newfoundland and Labrador came under fire in 2025 for publishing an education report and health human resources plan created by Deloitte, both containing fake citations likely generated by AI.

Is the GNWT’s guideline enough?

One NWT government employee, who spoke to Cabin Radio on the condition of anonymity to speak freely about their employer, said the territory’s generative AI guideline contains “just suggestions” and they are concerned about the lack of more comprehensive AI policies.

“Obviously governments are going to look into technologies for increasing efficiencies, but integrating AI into workflows is not as simple as giving staff access to a new app and giving them some ‘dos and don’ts.’ But that’s how it’s been rolled out,” they said.

The employee said existing security, privacy and electronic information policies were not written with AI in mind. They believe those policies are not enough to protect sensitive information, nor ensure oversight and accountability related to the use of AI within the GNWT.

“As far as I’m aware, there are no AI experts employed at the GNWT who could ensure AI programs are working properly or are producing unbiased content,” they said, adding it is unclear who is providing AI oversight in each department.

Andrew Fox, the NWT’s information and privacy commissioner, said the GNWT having a generative AI guideline and thinking about its approach to AI is good.

“It’s definitely an important sphere of activity to be paying attention to and the fact that … they’ve rolled out this high-level guideline is very encouraging,” he said.

But Fox said the guideline is “more of a statement that there should be policies and guidelines, as opposed to actually being sort-of applicable policies or guidelines for any particular AI application.”

Fox said he does not “think there is any question” the GNWT needs to develop further policies and guidance related to specific AI tools.

“It does seem to me that you can’t have sort-of one policy to rule them all,” he said.

‘Unfocused and unclear’

Teresa Scassa, Canada research chair in information law and policy and a faculty member at the University of Ottawa, said “it’s really important” that governments have policies regulating employee use of AI.

Scassa said the GNWT’s guideline is not adequate to address potential issues with AI use.

She described the guideline as “a little bit unfocused and unclear,” saying it muddles together the use of AI programs by departments and agencies as well as ad-hoc employee use.

She said it also does not lay out the chain of responsibility for decision-making and does not specify that employees can only use AI tools that have been vetted and approved by the government.

“The policy seems to say that it wants to encourage responsible use within government to tap the potential or benefit of it,” she said. “But then who’s making those decisions? Who’s responsible for those decisions and what do those people who are responsible need to do to put [them] in place? It’s very unclear about that.”

Scassa said the GNWT should further consider culturally sensitive information related to AI use. She said that should include engagement and consultation with Indigenous communities.

Gayla Thunstrom, president of the Union of Northern Workers, said the union is concerned about the appropriateness of using AI to produce content that requires cultural sensitivity or awareness.

“NWT residents deserve culturally appropriate public services, delivered by people,” she said in a statement.

Thunstrom said while some types of AI that can assist workers “may be worth consideration,” the union does not support any technology replacing people or bargaining unit work, or being used to fill vacancies.

She said the union would like more information on how errors or omissions related to AI use could impact employee reviews, as well as on models that are influencing any programs used by GNWT employees.

Thunstrom added the union has questions about the environmental impacts of AI, such as how the GNWT plans to mitigate its carbon footprint and whether it will be including data on AI use in greenhouse gas emissions reporting.

AI policies elsewhere in the NWT

The City of Yellowknife currently has no formal guidelines or policies governing the use of AI by employees but said it is planning to develop guidelines this year.

The municipality said access to AI tools at City Hall is currently limited to safe web services and employee use of AI tools is subject to existing policies related to privacy, records management and cybersecurity.

Aurora College president Angela James said the college has developed guidelines for the use of AI that cover students, staff, faculty and contractors. She added that statements and guidelines on the use of AI are included in course syllabi.

The Law Society of the NWT released guidelines in January 2025 for the use of generative AI in the practice of law. The guidelines state that lawyers must uphold professional responsibility standards and ensure compliance with the code of professional conduct while “embracing new technologies.”

The Supreme Court of the NWT also issued a notice to the public and practitioners in October 2025 urging caution when using AI.

The court stated that AI-generated court submissions must be verified and people need to inform the court if they submit documents that include AI-generated content.

Related Articles