Brief to the Federal Government Artificial Intelligence Strategy Taskforce October 2025

Our first recommendation to the Strategy Task Force involves clarifying the federal political leadership role on AI. As of now, the federal government has at least 4 Ministers involved with AI issues from Evan Solomon, Minister of Artificial Intelligence, Treasury Board President Shafqat Ali, Joel Lightbound, Minister of Government Transformation, and finally Patty Hajdu, Minister of Labour. But the roles of each of them regarding AI is not clear. We think that the respective responsibilities of each vis a vis AI should be clarified.

Our second demand, which has been talked about but not yet implemented, is a public registry with immediate registration of new and planned AI projects. This registry should be public and on line and contain the effects in terms of any job changes, losses or gains due to use. We would also like Statcan to survey these changes on a regular basis and make these results available.

The presence of unions and community organizations should be mandatory on all government AI bodies. For example, the AI Advisory Committee has never had either union or community representation and has been limited almost solely  to science (Academic) and business representation. But AI deeply affects both labour (working Canadians)  and community in a myriad of ways and it is important for the government to hear and consider the views of working people’s and community organizations as to how AI touches their lives.

The Government has promised a new bill on AI replacing C-27. C-27 had many flaws which need to be rectified in a new Bill. We support a new bill but:

1. The Bill must have a long period of  public consultation before a House of Commons Committee which will allow all organizations and all sectors which wish to contribute to the design of the Bill.  The last time no such important consultations were held and many sectors and organizations were excluded from presenting their views .
2. The new bill must also include AI Regulation for the Federal Government and federal crown corporations both of  which were excluded from C-27. The Government cannot limit regulation only to the federally regulated private sector but must include the wider private and public sectors.
3. The new Bill must also include a new Independent AI Regulator on all issues which citizens, governments, non-governmental organizations  and companies experience with AI.  In the radio, television, telephone and internet services, Canada has the CRTC as an independent regulator. We need an equivalent AI body.

While we have a Privacy Commissioner we need particular protection on the use of AI to monitor citizens and workers and to control data generated by AI. Employees and citizens must have informed consent on AI use and must be informed of any use of AI by governments or employers and have the right to refuse AI surveillance mechanism use. Also all citizens and all employees must have the right to inform the AI Regulator about AI activities which negatively affect them without suffering any job or other consequences.

All Employees affected by AI must have:

1)     adequate warning as to the timing of AI implementation in their workplace (a minimum of 180 days) and impact assessment results of the consequences of this AI implementation.

2)     also warning  if and how algorithmic assessment is being used to manage their jobs.

3)     access to  and funding for job training to allow them to adapt to AI use in their existing jobs.

4)     guaranteed protection against potential layoffs due to AI and that must involve:

•       the right to training for new job opportunities
•       funding for possible moves to new job locations in another part of the country
•       if none of these above possibilities work, workers who are laid off must have access to improved EI and pension programs which may need to be speeded up for workers near retirement age.

1.     Collective Bargaining Rights on AI  

Workers and their unions must have the right to bargain AI issues in their collective agreements. AI use cannot be seen as an exclusive management right.

2.   AI in the TBS and Government Agencies

Most PIPSC members work for the federal government and its agencies and many are engaged in developing or using AI.  In March 2025 the Treasury Board brought in an AI Strategy.

1) In the process of this unveiling, the TBS leadership promised there would be a TBS union meeting on AI. But this has yet to occur. We think establishing a permanent federal government and unions AI committee would be an essential move to assure AI is developed and used in the federal government in the best possible manner.

2) All AI decisions reviewed by humans (rescind sections in Directive on Automated        Decision-making)

In the Directive on Automated Decision-making, decisions in the federal government are divided into 4 and 2 of the 4 types of decisions do not need human review. We think this is wrong and that all decisions must have human review.

3) TBS and government agencies must prioritize the use of civil servants to develop AI in the government and avoid except in extremely exceptional circumstances of using outside consultants.

Canada must develop and use in priority its own Canadian owned AI resources. Chatbots like Open AI’s Chat GPT. Microsoft’s Copilot or Google Gemini are all US owned as are the major data centres where much of Canada’s data is stored. We need to build new Canadian owned data centres and AI systems and store our data 100% in Canada.