AI TOOLS REQUIRE EXPERT OVERSIGHT TO AVOID LEGAL RISK

AI as a tool contributes to faster and more efficient outcomes. It finds, writes, compiles and generates documents and images based on your prompts. But that is key – your prompt.

Think of your prompt as an instruction. The more expertise you have, the better you can prompt AI and evaluate its output. If AI is a tool in your workplace, it is likely performing work alongside many of your employees. Therefore, its results are likely being incorporated into your work products. How that happens could affect the quality of your work, which may impact how fast your business grows or legal risks it encounters.

Responsible use of AI in your workplace

You are the expert, not AI

AI output needs to be verified and this is quite difficult to do when you have only introductory knowledge on any given topic. Experts will more quickly and easily identify hallucinations, improper context or incorrect information, so the end product will be more reliable.

AI results need to be checked

Currently, AI tools are prone to many issues that if not caught early, could lead to exponential expansion of falsities, if others are relying on the output or incorporating it into other work products. Consider all AI results as plausible, rather than truthful. Experts will carefully correct AI outputs.

Know your tool

Being aware of how AI pulls in its sources, or answers your questions allows you to know the limits of what it can do. This is extremely important when its outcome is influencing decision making. AI results are starting points, not conclusive information. Experts need to consider both AI’s findings as well as anything outside the scope of AI results.

Be specific

AI will answer your questions but are the answers applicable to your exact scenario? If answers may change in different jurisdictions, did you provide that information in the prompt? Just because you asked, and AI answered, doesn’t mean that answer is relevant. Experts will give AI detailed and specific prompts.

“Don’t Judge a Book by its Cover”

Just because AI gave you an answer that looked reliable, with sources and citations, etc. It could very well have made those up, or attributed a quote that didn’t exist, or created new facts or assumptions. Experts will verify these details.

AI Tools might breach privacy

Most consumer level AI products do not offer privacy protections for personal information, but enterprise level products usually do. If your company collects personally identifying information such as client names, contact Information, birth dates or sensitive information such as financial or health data, it’s quite important to ensure your tech stack works with you to protect this data. Regulated professionals are professionally obligated to do so. Breaches of privacy are costly mistakes. One way to mitigate is to use contract agreements with your vendors and outline how data is handled and where it is stored.

How does AI “Think”?

As Sarah Hirebet, Director of Knowledge and Innovation, at Stradley Ronon, notes at a recent CLIO presentation during the Legal AI Accelerator Summit in August 2026,

Large Language Models “LLM” are designed to:

  1. Identify statistical patterns from their training data.
  2. Future answers may pull from incorrect training data, so AI might ‘believe’ a false input and give that answer to someone else.
  3. Reproduce biases based on its training input.
  4. Predict words, not retrieve facts; it is not a search engine.
  5. Make profit.
  6. Tell you something positive about any question you ask.
  7. Provide ‘likely’ outputs rather than truthful results.

RESPONSIBLE USE OF AI IN YOUR BUSINESS CAN PROTECT YOUR COMPANY

Businesses have many intangible assets, such as your goodwill or reputation and when these are harmed, business may suffer. Recovering may take a long time, as it’s not easy to rebuild trust with your clients. Companies that take AI and its use in the workplace seriously can proactively mitigate harm that may arise with irresponsible use of AI.

The best way forward is to develop an AI Use policy and provide training on incorporating AI tools in your company. Policies ensure all your employees know what is expected when using AI.

COMPLIANCE IS NOT AN OPTION, AND EMPLOYERS ARE RESPONSIBLE FOR DECISIONS WHETHER THEY WERE MADE OR INFLUENCED BY AI

Businesses that understand the benefits and limitations of AI will best be able to develop useful AI Policies for the workplace, that protects the company from preventable and irresponsible mistakes. Currently, employers with more than 25 employees need to disclose when AI is used to screen assess or select applicants during the recruiting process [*see our blog post AI Hiring Tools: Here’s How Employers Can Stay on the Right Side of the Law*], but the general use of AI in businesses is not regulated in Ontario. However, businesses still must adhere to privacy, human rights and employment laws, so developing policies to ensure compliance is a wise business decision.

[see our blog post on Why Every Employer Now Needs an AI Policy]

How SpringLaw can help

Springlaw is a tech focused employment law firm and can assist in developing Responsible Use of AI Policies, governance structures, compliance frameworks, and vendor contract agreements that protect your organization while allowing employees to benefit from responsible AI-driven productivity gains. Contact Us!

Even if your company hasn’t formally adopted AI yet, you still need an AI policy. This is true even if your company doesn’t allow employees to use AI or if it isn’t planning on purchasing any AI tools.  

The reality is, if you employ people, AI is already present in your workplace. The only question is whether you are managing it or ignoring it.  

“We Don’t Use AI” Is Usually Not True 

Most employers don’t realize how often AI shows up in daily work. 

Employees can use AI for a variety of purposes, including: 

  • Draft emails and reports 
  • Summarize meetings 
  • Generate presentations 
  • Screen resumes 
  • Prepare job descriptions 
  • Research legal, technical, or operational issues 

They can do this using free, public tools that sit outside your systems and controls. 

In fact, a 2025 study of more than 32,000 workers from 47 countries found that nearly 3 in 5 employees use AI intentionally and regularly at work. The study covered all geographical regions and occupational groups. This means that even if leadership never approved AI use, it is likely still happening, and without a policy, it is happening without guardrails. 

The Legal Risk of Silence 

When there is no policy, employees can make their own assumptions. Some may assume that everything is allowed, while others may assume that nothing is allowed and choose not to disclose their use of AI. Both scenarios create risk for the company. 

Silence can create risk around: 

  • Confidential information 
  • Intellectual property ownership 
  • Accuracy of work product 
  • Accountability for errors 
  • Compliance with privacy obligations 

AI Policies Are Not About Banning Technology 

A good AI policy is not about fear or control. We’ve written previously that companies should avoid bans on AI, since a likely effect would be to drive AI use underground. The benefit of a clear AI policy is that you bring AI use into the open, where it can be discussed and regulated. 

An effective policy helps employees understand: 

  • What tools are acceptable 
  • What information must never be shared 
  • When AI can and cannot be used 
  • Who remains responsible for final work 

Confidential Information Is a Major Exposure 

One of the most common AI mistakes employees make is sharing confidential information with public tools. 

Confidential information can include: 

  • Client data 
  • Employee information 
  • Financial details 
  • Business strategies 
  • Draft contracts or policies 

Once that information is entered into a public AI tool, employers may lose control over it. It may even be stored in another country, and, if an AI engine trains on it, it may expose your confidential information to the public, which creates privacy risks and potential contractual breaches. 

This is why an AI policy should clearly state that confidential information must not be entered into public AI tools. It should also clarify what kind of information is confidential. 

Intellectual Property Questions Employers Overlook 

AI also raises ownership issues many employers have not considered. 

If an employee uses AI to create content, who owns it? What if the employee entered company intellectual property into an AI tool, is it still protected?  

Without a policy, employers may face disputes over: 

  • Ownership of AI-assisted work 
  • Use of AI-generated materials 
  • Infringement of others’ intellectual property 

Clear rules help avoid messy arguments later. Make sure your AI policy clearly states that none of your company’s intellectual property can be entered into a public AI tool. 

Hiring and HR Are High-Risk Areas 

AI policies also matter when it comes to hiring and people management. 

As discussed in earlier blogs, AI hiring tools introduce human rights and disclosure obligations. Policies help ensure: 

  • Only approved tools are used 
  • Human review and decision making are required 
  • Bias and discrimination risks are monitored 
  • Disclosure requirements are met 

Policies Support Consistency and Fairness 

One of the strongest arguments for an AI policy is consistency. Without a clear policy, there is a risk that every department or employee could use AI differently, resulting in a variety of different risks to the company that may not come to light until it’s too late. A policy creates a shared baseline for everyone. 

Training Makes Policies Work 

A policy that sits unread in a handbook does very little. Employers should pair AI policies with basic training that explains: 

  • Why the policy exists 
  • How AI tools work at a high level 
  • Common mistakes to avoid 
  • Where to ask questions 

In particular, before adopting a specialized AI tool, such as one for hiring, companies should ask the AI provider if there is training available. 

It’s also advisable to arrange for AI training for employees on a regular basis, or to provide regular time for self directed learning, in order to ensure their skills are up to date. 

What a Strong AI Policy Should Cover 

Every workplace is different, but most AI policies should address: 

  • Approved and prohibited uses 
  • Confidentiality and privacy 
  • Accuracy and verification 
  • Human oversight 
  • Accountability 
  • Compliance with employment and human rights law 
  • How violations are dealt with (discipline) 

Waiting for a Problem Is the Expensive Option 

Many employers only address AI after something goes wrong, such as a data leak, a complaint from a third party, or a public mistake. At that point, the conversation is more about damage control than prevention. Policies are far easier to implement in the prevention stage, when you have the time to think and discuss with your team about how you want AI used in your company. 

The Bottom Line for Employers 

Whether you are aware of it or not, your workplace is already interacting with AI in some way. Ignoring that reality increases legal and reputational risk. An AI policy doesn’t need to be long or complicated, but it should be clear, practical, and aligned with how your people actually work. 

If you want help drafting an AI policy that fits your business, reviewing current practices, or training teams to use AI responsibly, contact us, call us, or schedule an appointment with us. 

Psychological safety has evolved from an HR buzzword into a recognized area of legal risk. In Ontario, employers are increasingly expected not only to prevent physical injuries but also to take reasonable steps to protect workers from psychological harm. Mental health in the workplace is not an issue employers can afford to treat lightly. When these obligations are overlooked, the consequences can surface in claims before the Workplace Safety and Insurance Board (“WSIB”), applications under the Ontario Human Rights Code, or other costly and time-consuming legal disputes. 

So, what does psychological safety mean in practice, and where does the WSIB fit in?  

Psychological Safety: More Than Being “Nice” 

While not formally defined in Ontario legislation, psychological safety is commonly understood to mean that employees can speak up, raise concerns, and do their work without fear of humiliation, retaliation, or emotional harm. Ontario recognizes that workplace factors can contribute to psychological injuries, including conditions like post-traumatic stress disorder, and that employees may face discrimination based on mental health disabilities or addictions.1  

Employers are increasingly expected to proactively identify and manage psychological hazards in the workplace. These hazards can include bullying, harassment, excessive workloads, unclear roles, and toxic leadership. Meeting this expectation requires concrete actions: assessing risks, providing training, and taking employee concerns seriously. 

These responsibilities are reinforced through overlapping legal obligations under the Ontario Occupational Health and Safety Act, the Ontario Human Rights Code, and policies administered by the WSIB. In practice, this means it is in employers’ best interest to take tangible steps to prevent psychological harm and document their efforts to comply with legal requirements and recognized best practices.  

When Psychological Injury Becomes a WSIB Claim 

This is where things get real. Ontario’s Workplace Safety and Insurance Act allows workers to claim WSIB benefits for diagnosed work-related mental stress injuries, including, but not limited to: 

  • Chronic mental stress, where a work-related stressor is the predominant cause, such as workplace harassment. 
  • Traumatic mental stress, caused or significantly contributed to by one or more work-related traumatic incidents. 

These categories are distinct. To qualify for chronic mental stress, the worker must have a properly diagnosed mental stress injury caused by a substantial work-related stressor. “Substantial” means excessive in intensity and/or duration compared to normal workplace pressures.2 In other words, this is not simply about a bad day with a difficult manager. 

Workplace harassment, bullying, or egregious conduct often meets this threshold. Yelling, demeaning comments, threats, or sustained verbal abuse may qualify when they go beyond ordinary interpersonal conflict and create a toxic or abusive work environment.  

What WSIB Usually Won’t Cover 

Not every stressful workplace experience leads to entitlement. WSIB generally excludes mental stress caused by legitimate management actions that are part of the employment function, such as discipline, performance management, changes to duties, or termination, even when those decisions are naturally upsetting. 

Interpersonal conflict also typically falls outside WSIB coverage unless it rises to the level of harassment or objectively abusive conduct that a reasonable person would view as egregious. The focus is on what happened, not just how it felt. 

Ontario applies a “thin skull” approach: a pre-existing mental health condition does not automatically bar entitlement if the workplace conduct significantly contributed to the injury. WSIB decisions are ultimately made on a case‑by‑case basis.  

Why This Matters for Employers 

Psychological safety failures don’t just affect morale; they can trigger WSIB claims, investigations, and expensive disputes. Employers are expected to act once they know, or reasonably should know, that there is a workplace problem, even if no formal complaint has been made. 

The upside? Strong psychological safety practices are also smart risk management. Clear policies, meaningful training, prompt investigations, and well-documented responses go a long way not just toward legal compliance, but toward healthier and more resilient workplaces. 

In Ontario, psychological safety is no longer optional. It’s part of the legal landscape and well worth employers’ attention. 

SpringLaw helps employers design practical policies that reduce legal risk while supporting a healthy workplace culture.  Contact us if you need assistance developing policies or procedures to protect the psychological safety of your employees and reduce your legal risk. 

Constructive dismissal arises when an employer makes a significant change to an employee’s terms of employment without their consent. Even where the employee hasn’t been expressly terminated, the law may treat the change as a termination if it fundamentally alters the employment relationship. 

Common examples include: 

  • Reductions in compensation; 
  • Changes to job duties or reporting structure; 
  • Relocation or changes to work location (including return-to-office mandates); or  
  • Demotions or changes in status. 

For employers, the challenge is that businesses need to evolve. Roles change, structures shift, and operational decisions need to be made. The legal risk arises when those changes cross the line from reasonable management decisions into fundamental changes to employment terms. 

There isn’t a bright-line rule, but the analysis often turns on the degree of change and whether the employee agreed (explicitly or implicitly). 

So how do employers balance this? 

A few practical considerations: 

  • Assess whether the change impacts a core term of employment; 
  • Consider whether the employment agreement allows for flexibility or change; 
  • Provide advance notice where possible; 
  • Communicate the business rationale clearly; and  
  • Consider obtaining consent to the change.  

In some cases, employers may also consider providing working notice of the change (effectively treating it as a termination and offering continued employment under new terms after the notice period). 

Constructive dismissal risk doesn’t mean change is off the table. It does mean that how a change is implemented really matters. A thoughtful, transparent approach can help reduce risk while still allowing the business to move forward. 

The right approach to change can protect both your business and your workplace relationships. If you’re unsure how a proposed change may impact your obligations as an employer, we’re here to help.

You are an employer in Ontario. An employee has raised concerns about a colleague’s behaviour. Whether the allegations could amount to workplace harassment, workplace violence, discrimination, or other inappropriate conduct, your response matters.

It’s natural to have an immediate reaction. Pay attention to that!  You may believe the respondent would never act that way, or you may feel the complainant has a history of raising concerns. However, when workplace allegations arise, assumptions can quickly create legal risk. Ask yourself: if your first instinct is to dismiss the complaint, or if those around you are inclined to downplay it, that’s a sign. The safest approach is to pause, remain objective, and follow a fair investigation process. It may be time to consider bringing in an external, independent investigator. An independent investigation can help ensure the process is objective, credible, and aligned with legal requirements.

Many employers are surprised to learn that a workplace investigation begins long before interviews are scheduled. Key decisions about communication, documentation, interim workplace measures, and confidentiality can significantly affect the integrity of the process. Getting guidance early can help employers navigate these challenges while maintaining trust and fairness for everyone involved.

Over the course of my career, I’ve not only conducted workplace investigations as an investigator, but I’ve also guided employers, complainants, and respondents through the investigation process. One thing I’ve seen repeatedly is that organizations often struggle most at the beginning. Determining how – and how quickly – to respond, what steps to take, and whether an external investigator is needed can be just as important as the investigation itself.

Ontario law has evolved, and employers now have a clear obligation under the Occupational Health and Safety Act (OHSA) and the Ontario Human Rights Code. When allegations arise that could amount to harassment, workplace violence, or discriminatory conduct, employers must conduct an investigation that is reasonable and appropriate to the situation. Confidentiality is key. During the process, it’s crucial to protect the privacy of all parties involved. Ontario law also prohibits retaliation. No one can face negative consequences for participating in good faith.

Small to medium-sized employers are sometimes tempted to address workplace complaints informally or internally without a structured process. But under the OHSA and the Human Rights Code, that approach can create risk. A reasonable, appropriate, and neutral investigation is not just best practice. It’s the law.

From an employer’s perspective, a well-managed workplace investigation is about more than legal compliance. It helps demonstrate that concerns are being taken seriously, supports psychological safety in the workplace, and can reduce the likelihood of further conflict or litigation.

In today’s legal landscape, ignoring or brushing off allegations can open the door to legal liability. By understanding your duties under Ontario law and recognizing when to seek independent investigation support or guidance through the process, you can better protect your organization, reduce legal risk, and foster a safer, fairer workplace for everyone.

For Ontario employers, taking every complaint seriously and following through with a fair, appropriate investigation isn’t optional. It’s essential. Contact us for more information and guidance.

Technology has made it easier than ever to monitor how work gets done. With the rise of hybrid and remote work, digital productivity tools, and AI‑enabled systems, workplace monitoring has become a key compliance issue for Ontario employers.  

Ontario does not have a single, comprehensive statute governing employee privacy in the workplace. However, under the Employment Standards Act, 2000 (“ESA”), employers with 25 or more employees in Ontario must have a written policy on electronic monitoring in place. This requirement was introduced through Bill 88, Working for Workers Act, 2022 and continues to apply each year. 

A Refresher on Who Needs a Policy and When 

If you employed 25 or more employees on January 1 of any year, you must have a written electronic monitoring policy in place before March 1 of that year. This headcount includes anyone who meets the definition of “employee”, including full‑time, part‑time, casual, and seasonal employees. 

Even if you do not engage in electronic monitoring at all, you still need a policy that clearly says so. 

What Must the Policy Include 

The ESA is focused on transparency, not prohibition. A compliant policy must state: 

  • Whether the employer electronically monitors employees; 
  • How and in what circumstances monitoring occurs; 
  • Why the information is collected and how it may be used; and 
  • the date (including day, month, and year) the policy was prepared and last updated. 

Employers must also provide a copy of the policy to employees within the required timelines. 

Importantly, the ESA does not create new privacy rights or give employees a right not to be monitored. Nevertheless, that does not mean employers have free rein. 

What Should Employers be cautious about 

Even though Ontario lacks a standalone law governing employee privacy in the workplace, privacy expectations still matter. Courts and arbitrators consistently assess whether monitoring is reasonable and proportionate, and whether it serves a legitimate business purpose. 

Common risk areas include: 

  • Overly broad monitoring tools that capture personal use; 
  • Monitoring that is not clearly disclosed in the policy; 
  • Collecting more information than is necessary; and 
  • Using monitoring data for purposes not described in the policy. 

In a grievance between Rideauwood Addiction and Family Services and the Ontario Public Service Employees Union, the arbitrator held that the key issue was whether the electronic monitoring policy and the monitoring it permits constituted a reasonable exercise of management rights in all of the circumstances.1 

While monitoring work‑related communications may be reasonable, a policy that broadly monitors internet and network activity without distinguishing between work-related and personal use, and without adequate justification, may be found unreasonable. 

Workplace policies can reduce an employee’s expectation of privacy, but they do not eliminate it entirely. 

Key Takeaways for Employers  

  • Review your monitoring practices annually against the ESA requirements. 
  • Clearly distinguish between work-related and personal use in your monitoring policy.  
  • Avoid vague, “catch-all” language that allows unlimited monitoring. 
  • Make sure your policy accurately reflects what you actually monitor in practice. 
  • Ensure your monitoring is justified and proportionate. 
  • Regularly review policies to ensure they remain legally defensible and aligned with evolving laws and best practices.  

A clear, carefully drafted policy can help manage risk while still giving employers the flexibility they need to operate their business. 

SpringLaw drafts privacy policies that balance compliance with practical business needs. Contact us anytime.  

  1. Ontario Public Service Employees Union v Rideauwood Addiction and Family Services, 2024 CanLII 120507, ONLA at para 22.

DECISION: Cillis v. Hamilton-Wentworth District School Board, 2025 HRTO 1717 

A decision of the Human Rights Tribunal of Ontario (“Tribunal”) explores the boundaries of the duty to accommodate in the workplace, raising important questions about how far employers must go to meet accommodation requests, and whether emotional discomfort during the process can establish a basis for a discrimination claim. This decision provides clarity on the procedural and substantive dimensions of the duty to accommodate, reinforcing that employers are not obligated to meet every preference or expectation. 

Facts 

The applicant was employed by the Hamilton-Wentworth District School Board (the “School Board”), and she participated in an interview process for a permanent teaching position. As a result of a mental health disability, the employee provided documentation from a healthcare provider stating that the employee was “able to recognize many of the triggers that exacerbate her condition, and this should be heeded to avoid any emotional distress to her mental health.” The employee also requested accommodations for the interview process, such as a private space to relax before the interview, additional time to review interview questions, and the exclusion of a specific individual from the interview panel. 

The School Board provided most of the requested accommodations, including removing the individual from the panel, offering a private space, and providing more preparation time than other candidates had received. The employee was ultimately successful in the interview process, and she was offered a permanent position. However, the employee alleged that the accommodations were deficient and caused her emotional distress. She cited less preparation time than she had expected (30 minutes instead of the expected 60 minutes), comments made to her that she found dismissive and anxiety-inducing, and the presence of individuals in the hallway (including the removed interviewer). 

Legal Framework 

The Tribunal applied the established three-part test for prima facie discrimination: 

  1. The applicant must have a protected characteristic under the Human Rights Code (“Code”); 
  2. The applicant must have experienced adverse treatment; and 
  3. The protected characteristic must have been a factor in the adverse treatment. 

While the employee’s disability was undisputed, the Tribunal found that the second and third elements of the test were not met. 

Although the employee argued that the process lacked sensitivity and dignity, the Tribunal found that the accommodations provided were reasonable and met the legal standard. The Tribunal clarified that the Code does not require employers to meet every subjective expectation. The Tribunal acknowledged that the accommodations were not exactly what the employee had expected; however, the Code only requires “reasonable” accommodation as opposed to “perfect” or “ideal” accommodation. 

The Tribunal also noted that the employee’s own physician had not specified any strict limitations or requirements beyond general recommendations to avoid emotional distress. The accommodations provided were consistent with the medical documentation and the employee’s stated needs. Although the employee described negative feelings and feelings of anxiety during the accommodation process, the Tribunal found that those feelings did not amount to adverse treatment or substantive disadvantage sufficient to establish a prima facie case of discrimination. 

Takeaways 

This decision serves as a reminder that the accommodation process must be respectful and responsive, but employers are not bound by the subjective preferences of the employee. This case makes clear that emotional discomfort may not be sufficient to establish discrimination under the Code, in the absence of any substantive adverse impact or disadvantage. It is a valuable precedent for understanding the limits and obligations of accommodation in the workplace. 

If you have questions about how this decision might apply to your workplace, find yourself dealing with a similar situation, or would simply like guidance on medical leaves and accommodation obligations, the team at SpringLaw would be happy to help. We regularly advise both employers and employees on navigating these issues in a practical and strategic way, feel free to reach out to discuss your specific circumstances. 

Most conversations about AI at work focus on which tools to adopt and what policies to put in place. Those things matter, but they are not actually where the hard work happens. The harder work is getting your team into the right headspace to engage with change in the first place, and that is a culture question, not a technology one. 

Mindset is the bottleneck 

The pace of change with AI is genuinely fast, and it is not slowing down. Most teams are not struggling because they lack access to good tools. They are struggling because some people feel curious and energized by what is new, while others feel unsettled or sidelined by it. That gap, between the people who lean in and the people who pull back, is where organizations lose momentum. 

How people respond to change is shaped heavily by the culture around them and by what leaders model day to day. That means it is something you can actually work on. 

What a growth mindset actually looks like on the ground 

A growth mindset is not a personality trait and it is not a workshop you send people to once. It shows up in small, observable behaviours: asking questions instead of defaulting to skepticism, treating early stumbles with a new tool as information rather than failure, and staying genuinely open to the possibility that the way things have always been done might not be the best way going forward. 

One of the most effective ways to build this across a team is to make learning social and low-pressure. A few approaches that work well: 

  • A regular “show and tell” session where team members share what they have been experimenting with, what worked, what did not, and what surprised them. 
  • A shared channel or running document where people drop AI tips, useful prompts, or shortcuts they have discovered. This keeps learning visible and ongoing rather than siloed. 
  • Short lunch-and-learn sessions built around a single tool or use case, with a debrief at the end. The debrief matters as much as the demo. 
  • Pairing people up across experience levels so that someone more comfortable with a tool can work alongside someone who is less so, without it feeling like formal training. 

The point is to normalize experimentation as part of how work gets done, not something that happens off the side of the desk when there is spare time. 

Leaders go first 

If the owner or most senior person in the room is visibly skeptical or disengaged, the team reads that immediately. Culture follows behaviour, not intention. Leaders who want their teams to develop a growth mindset need to model it openly, which means saying “I tried this and it did not work the way I expected” in front of the team, asking questions publicly, and recognizing curiosity and effort, not just results. 

Address the fear directly 

Anxiety about AI is real and it is not going away on its own. Employees worry about relevance, about whether their skills will still matter, about what their job looks like in a few years. The organizations that handle this well talk openly about where AI is changing the work, what it is not replacing, and what the organization is doing to help people grow. That kind of transparency builds trust in a way that a policy document never will. 

When the fit is not there 

Some team members will resist regardless of how much support and encouragement they receive. The right response is to make sure the opportunity to engage has been genuinely offered through training, peer learning, and clear conversations about expectations. After real time and real effort on both sides, some people will still not be the right fit for where the organization is heading. That is a legitimate outcome, and addressing it honestly is better for everyone than letting it drag. 

Building this takes time, but starting is the point 

A team with a genuine growth mindset is one of the most durable advantages an organization can have as AI continues to change how work gets done. It is built through consistent leadership behaviour, a culture that rewards curiosity, and a willingness to have honest conversations when things are not working. 

If you are thinking about how to build this in your own organization, or navigating a situation where resistance has become a real issue, we can help. Reach out at springlaw.ca.

Ontario’s introduction of long-term illness leave (“LTI Leave”) significantly expands statutory leave entitlements. While the statutory framework appears straightforward, the real legal risk lies in how it intersects with human rights accommodation obligations. 

The Statutory Framework 

Effective June 19, 2025, the Employment Standards Act, 2000 (“ESA”) provides eligible employees with up to 27 weeks of unpaid, job-protected leave in a 52-week period due to a serious medical condition.  

To qualify, employees must: 

  • Have been employed by their employer for at least 13 consecutive weeks  
  • Provide a medical certificate from a qualified health practitioner confirming a serious medical condition and the period of the employee’s inability to work  

Employers must: 

  • Accept reasonable medical evidence (not detailed diagnoses)  
  • Hold the employee’s position  
  • Continue benefit plans  
  • Reinstate the employee to their former position, or a comparable role if it no longer exists 
  • Ensure the employee’s pay is not reduced because of the leave 

Failure to comply may result in ESA complaints, penalties, and orders for reinstatement. 

Where Employers Face Greater Risk: Human Rights Law 

Compliance with the ESA does not end an employer’s legal obligations regarding disability-related absences. The Ontario Human Rights Code (“Code”) imposes a broader duty on employers to accommodate employees with disabilities to the point of undue hardship.  

Courts and tribunals have emphasized that ESA leave entitlements are minimum standards and do not replace the separate, individualized accommodation analysis required under the Code. The Supreme Court of Canada has confirmed that the duty to accommodate is measured by undue hardship, not “impossibility,” and that an employer’s duty ends only when the employee is no longer able to fulfil the basic obligations of the employment relationship in the foreseeable future.1 Ontario decisions likewise underline the need to canvass real accommodation options, such as modified duties or placement into available, suitable alternative positions, rather than treating a leave period as a hard stop.2  

The accommodation process is highly individualized and may involve altered duties, flexible work arrangements, or extended leave, provided these do not impose undue hardship.  

Treating the 27-week ESA illness leave as a hard cap, or terminating employment at the end of a statutory leave without a thorough accommodation analysis may expose employers to liability under the Code.  

Common Employer Missteps 

Employers most often face risk by: 

  • Treating LTI Leave as the maximum allowable absence 
  • Requesting overly detailed medical information  
  • Failing to document the accommodation process  
  • Not assessing graduated or modified return-to-work options  

Practical Takeaways for Employers 

  1. Update leave and accommodation policies to reflect LTI Leave.  
  1. Train HR and managers on the distinction between ESA leave entitlements and Code-based accommodation. 
  1. Track LTI carefully, but assess accommodation independently of ESA timelines. 

SpringLaw supports employers in managing illness‑related absences and human rights accommodation requirements effectively. Contact us today.  

AI is a useful tool for managing people in the workplace. It can track performance trends, flag patterns in attendance or productivity, and surface data that would have taken a manager hours to pull together. Used well, it makes day-to-day management faster and more informed. 

The problem is not the data. The problem is what happens when the data starts making the decisions. 

The gap between flagging and deciding 

There is a meaningful difference between AI flagging that someone’s output has dropped over three months and a manager deciding what to do about it. The first is a useful input. The second requires context, conversation, and judgment that a dashboard cannot provide. 

An employee’s performance may have declined because they are dealing with a serious personal situation, a health issue, or a dynamic on the team that the data does not capture. A system that sees a trendline has no way of knowing that. 

Why this creates legal exposure 

When an organization relies heavily on automated outputs to drive employment decisions, and those decisions are later challenged, the question that arises is whether a real and meaningful review actually took place. If a human was simply confirming what the system already decided, that is not oversight in any meaningful sense. It creates legal exposure and, frankly, it is not great management. 

Where human involvement is not optional 

There are specific moments in the employment relationship where a person needs to be genuinely in the driver’s seat. These include: 

  • Performance conversations that involve context or sensitivity 
  • Discipline processes that require judgment about proportionality and fairness 
  • Termination decisions, which carry enough legal and human weight that they should never be the output of an automated process 
  • Final-stage hiring decisions, where assessing fit and authenticity goes well beyond a scored profile 

None of this means AI has no role in these areas. It can help gather relevant information, structure documentation, or identify patterns worth investigating. What it should not do is own the outcome. 

What to do about it as a business owner 

The organizations that get this right tend to be deliberate about it ahead of time, not after something goes sideways. Here are a few practical steps to consider: 

  • Identify in advance which decisions require human sign-off, and make sure that sign-off is substantive rather than a formality 
  • Document your process, whether in a policy or internal guidelines, so there is a clear record of how decisions were made 
  • Train your managers to treat AI outputs as a starting point for a conversation, not a conclusion 
  • Review any AI tools you are currently using for workforce management and ask whether the outputs are influencing decisions in ways you have not explicitly approved 

Getting ahead of the gap 

If your organization is using AI tools for performance tracking, scheduling, or workforce analytics and you have not yet mapped out where human decision-making is required, that is worth addressing now. It protects your people and it protects you. 

If you want help thinking through where those lines should fall, we work through exactly these questions with employers regularly. Reach out at springlaw.ca.