Prompt potential, not risk
The AI prompts employees might be typing into online AI tools may create real risks for Canadian employers. A clear workplace AI use policy, paired with the right guardrails, may help reduce those risks and help workers use AI to enhance productivity and streamline basic tasks. Seemingly innocent prompts meant to save time may quickly lead to serious consequences.
This blog post will examine how everyday AI use by employees may create unnecessary risks for Canadian businesses. Learn from our HR experts about the privacy challenges a strong workplace AI use policy can address, plus actionable steps to help reduce risk and support your compliance efforts.
Using AI in human resources
Uploading employee records or company documentation into a public AI platform, asking ChatGPT or Copilot to screen a stack of resumes for a top candidate, or quickly generating a termination letter might seem like a great use of AI tools. Yet these uses could breach privacy law, introduce bias into hiring, or generate inaccurate case law and legal language that your organization can’t defend in court.
Which types of AI prompts carry the highest risks to employers?
Most prompts aren’t a major concern. The risky ones often follow a simple pattern: they feed the tool confidential information and ask it to make or validate a real-world decision. AI tools often miss critical context, overlook key details, or make assumptions based on incomplete information. That can lead to inaccurate, inappropriate, or non-compliant outcomes, especially when sensitive decisions are involved. Use AI as a support tool, not a final authority, and ensure careful human review before acting on its recommendations.
These are five common types of prompts to avoid:
1. “Attached is our employee handbook. Act as a Canadian HR professional and confirm that all required policies are compliant with current employment legislation.”
Without effective oversight, using AI tools can create serious compliance risks for Canadian businesses. These risks include accidental breaches of privacy laws, reliance on outdated or incorrect legislation, and AI producing inaccurate or misleading legal information. That’s why human-led review and clear policies are essential and may help reduce risk to your organization.
2. “Screen these 100 resumes I’ve attached and provide a list of the 5 best candidates for the position.”
Because AI learns from historical data, it can easily perpetuate existing biases. Before using AI to screen candidates, understand how it may impact fairness and bias in your HR processes.
3. “Write a termination letter for [employee].”
A termination letter needs to reflect the notice, pay, and other entitlements that apply in your jurisdiction and situation, and generative AI may miss these. Before acting, employers should contact Live HR Advice or obtain legal advice. Read our blogpost, How to Write a Termination Letter, with Samples, for a more comprehensive breakdown of what’s required.
4. “Draft a violence and harassment policy and cite the current law.”
Generative AI can misinterpret case law or invent interpretations that sound correct but actually aren’t. Any legal or legislative-heavy output is merely a draft for experienced human eyes to verify, not a ready-to-use policy to implement.
While AI can help you find Canadian laws, it won’t be held accountable if a decision doesn’t comply with employment legislation; your organization will.
5. “I’m conducting a workplace investigation. Here are the investigation notes and statements made by the parties involved; who’s telling the truth?”
Workplace investigations are some of the most complex and sensitive responsibilities HR professionals are tasked with, demanding careful consideration of context and human rights. Relying on an AI tool to assign responsibility or interpret details eliminates the nuanced, experienced judgment that HR or health and safety professionals provide.
The common thread across all five of these prompts is clear: an AI-generated answer that sounds confident isn’t necessarily accurate or defensible. Unlike HR professionals and managers, AI lacks personal relationships with employees and can overlook important context or mitigating factors.
What does Canadian legislation say about AI at work?
Privacy law should be your first consideration when using AI platforms. Depending on where you operate and whether your business is federally or provincially regulated, laws such as the federal Personal Information Protection and Electronic Documents Act (PIPEDA), Quebec’s Act respecting the protection of personal information in the private sector, and the Personal Information Protection Acts in British Columbia and Alberta may speak to personal information entered into an AI tool. Personal information entered as a prompt is still personal information, as is what the tool infers about a person.
The federal government’s guidance points in the same direction. Its Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems and the privacy commissioners’ Principles for responsible, trustworthy and privacy-protective generative AI technologies land on a shared idea: be transparent.
Keep employees involved at every stage. Never let AI make the final decision, and carefully control what personal or sensitive information is shared with AI tools.
What are the laws for AI in the hiring process?
The legal landscape for AI in hiring across Canada is fragmented. Most employment law is driven by the jurisdiction and privacy law is a mix of federal and provincial or territorial rules. While Ontario is the only province with an employment law specifically targeting AI in job postings, other provinces use strict privacy and human rights frameworks to regulate how AI handles candidate data.
AI in the hiring process across Canada:
| Jurisdiction | Legislation | Details |
| Ontario | Working for Workers Four Act, 2024 | Employers with 25 or more employees that use AI to screen, assess or select applicants must say so in each publicly advertised job posting. Employers must keep copies of each posting and any associated application form for three years. This applies whether or not AI is used. The requirements don’t apply to help wanted signs, internal hiring, or to positions where the work is performed outside Ontario. |
| Quebec | Law 25 — Act Respecting the Protection of Personal Information in the Private Sector | If AI makes a hiring decision with no human involvement, employers must inform the applicant. Candidates can request information about the personal information used and the reasons and principal factors behind the decision and submit observations to a staff member who can review the decision. Applicants must be told when technology that can identify, locate, or profile them is used. |
| British Columbia and Alberta | Personal Information Protection Act (PIPA): British Columbia and Alberta | Although not AI-specific, private employers must generally tell candidates before or at the time of collection why their personal information is being collected, including through AI tools. They must only collect what is reasonable for establishing an employment relationship. |
| Rest of Canada | No AI-specific hiring disclosure law | No AI-specific hiring disclosure law. Human rights laws apply, and privacy laws may apply depending on whether the employer is federally or provincially regulated. Nova Scotia proposed Bill 234, the Helping Working People Get Ahead Act, in 2026; however, it is in the early stages. Manitoba passed Bill 51, the Public Sector Artificial Intelligence and Cybersecurity Governance Act, which would establish a framework for regulating artificial intelligence in the public sector; however, the legislation is not yet in force. |
AI in hiring: Canadian legislation at a glance
- Ontario leads Canada in AI hiring law. Under the Working for Workers Four Act, 2024, employers with 25 or more employees that use AI to screen, assess, or select external applicants must say so in each publicly advertised job posting.
- Employers must keep copies of each posting and any associated application form for three years. This applies to all publicly advertised postings, whether or not AI is used, for employers with 25 or more employees. For a more comprehensive look at the requirements of the Working for Workers Acts, check out this blogpost.
- Quebec’s Law 25, The Act Respecting the Protection of Personal Information in the Private Sector, sets the privacy rules that affect AI hiring. If an employer uses AI to make a hiring decision without human involvement, they must inform the applicant. Candidates can also ask about the personal information used and the reasons and principal factors behind the decision. Also, they can submit observations to a staff member who can review them.
- The Personal Information Protection Acts (PIPA) in British Columbia and Alberta regulate how candidates’ personal information is collected and used, including through automated recruitment tools, though these provinces do not yet have AI-specific workplace legislation in effect. Private employers must generally tell candidates why their information is being collected, including through AI tools, and collect only what is reasonable.
- In the rest of Canada, no specific AI hiring disclosure laws are currently in force. Nova Scotia has a proposed act in the early stages and Manitoba has passed legislation establishing a framework for regulating artificial intelligence in the public sector, but it is not yet in force.
Remember, Canadian employers are still liable should AI tools lead to instances of discrimination, even without specific AI legislation in force.
How to turn AI into an advantage, not a risk
Many Canadian HR teams are already safely using AI to reduce administrative workload. This makes a clear workplace AI policy essential. Your workplace AI use policy should specify which tools staff may use, what information is prohibited from being entered into AI platforms, and when human approval is needed in the process.
If you manage HR alongside other responsibilities, don’t worry. You don’t need to become an AI expert. Rely on support from your IT people. Focus on providing your team with tools that streamline their daily tasks but always set clear rules to minimize risk.
Encourage staff to use AI for drafting and routine tasks but emphasize that they remain accountable for the final outcome. Also communicat that they must always adhere to company policies. Finally, ensure all AI use happens on secure, licensed business platforms and never on public, open-access tools.
Quick takeaways
- Keep personal information and sensitive company information out of public AI tools.
- Never let AI make the final call; hold humans accountable for the final outputs.
- Implement a clear workplace AI policy and training session before granting AI tool licenses.
Frequently asked questions
Do you need a workplace AI policy in Canada?
Legislation currently doesn’t require private-sector businesses to have an AI policy. But you’re still accountable for complying with privacy, human rights, and employment laws. A short policy that approves tools, specifies prohibited inputs, and requires human sign-off may help to reduce risks.
Is it legal to use AI to screen resumés in Canada?
It isn’t banned, but it carries risks. AI can repeat bias hidden in hiring data. Use AI to organize or summarize applications but keep someone accountable for the candidate shortlist and making the final decision. Employers should be able to demonstrate why particular candidates were selected or rejected. In Ontario, employers with 25 or more employees that use AI to screen, assess, or select applicants must disclose it in publicly advertised job postings.
Ready to use AI without prompting risk?
The businesses with the biggest AI wins aren’t the ones who wing it. They’re the ones who set the rules and keep humans involved in the processes from beginning to end.
Create a workplace AI use policy that helps protect your business and empowers your workforce
Start leveraging the real value of AI tools in your HR practices with content for every jurisdiction. With our HR, you’ll get:
- HR policies, templates, and checklists that support your compliance efforts across every jurisdiction.
- HR management software that uses AI and is backed by HR experts.
- Expert technical support.
Several of our packages also include live support from Canadian HR and health and safety professionals.
Book a demo today. See how our software, training, and live advice lines can help your team use AI for what it’s best at.