The federal government has unveiled its national strategy on artificial intelligence, committing over $2 billion in funding. But Prime Minister Carney’s strategy is silent on what workers actually need: guarantees AI won’t replace workers, surveillance protections, and public accountability. 

While the government boasts about responsible AI use on stage, the reality on the ground tells a different story.  

AI systems are already expanding across the federal public service, from deepfake detection at the Canada Revenue Agency to automated file-triaging at Employment and Social Development Canada. These systems are being deployed at lightning speed, and workers are being left behind. Employees asking for AI training are being turned down, and some managers are actively encouraging staff to use unvetted AI tools, even when it puts protected, sensitive data at risk. 

A poll commissioned by PSAC found that 56 per cent of people across the country don’t trust AI technologies to handle their sensitive data — including taxes, benefits, and pensions — and would rather deal with a human than an algorithm. The federal government’s own consultations on AI generated 11,000 submissions calling for risk-based regulation to ensure AI is used appropriately. 

Know your rights 

Management may be refusing to discuss AI, but the employer has existing legal obligations when it comes to technological change. 

Under Article 24 of PSAC’s federal collective agreements, management cannot simply add new software or automated systems into your regular workflows. They are legally required to provide at least 180 days' written notice and consult meaningfully with the union before making any changes that affect your job. 

What we're fighting for at the bargaining table 

PSAC is currently in negotiations with Treasury Board (EB, PA, SV, TC Groups), the Canada Revenue Agency, and Parks Canada, and our bargaining teams are pushing a historic set of new demands to protect members.  

Our proposals include: 

• Clear language stating that technological change must be used to augment — not replace — the work performed by employees 

• A full ban on using automated systems to make or support decisions on discipline, conditions of employment, or hiring 

• Restrictions on electronic surveillance or enhanced tracking to monitor performance, assess productivity, or discipline workers 

• Protection ensuring employees are never held personally liable for errors, inaccuracies, or unintended bias resulting from AI tools 

• Mandatory, ongoing training for all employees 

• A requirement that the employer prioritize building internal public service capacity instead of outsourcing to outside tech corporations 

• A dedicated National Joint Committee on Emerging Technology to review and manage the psychological health, safety, and privacy risks posed by AI 

When our Canada Border Services Agency (FB Group) and Canadian Food Inspection Agency bargaining teams exchange proposals with the employer later this year, they will also be pursuing similar protections. 

To be clear: PSAC is not opposed to technology. When implemented responsibly, with meaningful human oversight, AI can support the delivery of critical federal public services. But technology should never be used to cut corners, compromise privacy, or erode the human expertise and judgment that people across Canada depend on. 

We will keep fighting for a contract that protects our members' rights, safeguards public data, and ensures quality federal public services for everyone.