AI may make a job obsolete but it doesn’t eliminate an employer’s legal obligations

The most consequential employment decisions of the AI era may not be which jobs to eliminate but how to redesign the workforce.

Artificial intelligence is changing not only how employees work but also what employers can reasonably expect them to accomplish, which skills they require and how many of them are needed to run a business.

Tasks that once justified full-time positions may soon occupy a mere fraction of someone’s day. Entire departments may be reorganized around capabilities that did not exist when their employees were hired.

For employers, the commercial imperative is obvious; the legal implications, less so.

Employment law was developed around a stable premise: an employee accepts a defined role with specified responsibilities and compensation in exchange for performing work under the employer’s direction. AI is making the boundaries of that bargain increasingly murky.

The companies that navigate this transition successfully will understand a distinction that is easy to overlook: The right to transform a business does not include the right to transform employment terms or relationships.

Consider an employee hired as an analyst. The role involves gathering information, preparing reports, interpreting results and advising management. AI can now perform much of the preliminary work in minutes rather than days.

The employer can retain the employee and expect greater output, redesign the position around judgment and strategy, combine responsibilities previously divided among several employees or eliminate the role altogether. Any of these decisions are commercially rational. None are without legal consequences.

An employer that substantially changes an employee’s responsibilities, status or compensation may face a constructive dismissal claim, depending on the circumstances and the employment agreement. Introducing new technology does not permit the employer to rewrite the employment bargain.

The challenge is distinguishing between changing how work is performed and changing the terms on which an individual agreed to perform it.

If an employee can complete in two hours what previously took two days, should the employer expect 10 times the output? My legal answer is yes, and their performance can be managed accordingly.

But what if AI completes routine assignments brilliantly but produces errors on complex ones? What if one employee adapts readily to AI while another, equally experienced, struggles?

These questions affect performance reviews, promotions, compensation and dismissals.

Employers can change how work is organized and establish legitimate performance expectations, subject to applicable law and contractual obligations. But they must still be able to explain what they expect, how performance is measured and why an employee failed to meet the standard.

As with any other tool, there should be no expectation that AI performs every task reliably, that its output requires no serious verification or that every employee can use it equally effectively.

Nor should a company confuse an employee’s failure to produce more with a failure to adapt. The distinction requires evidence.

Where performance concerns intersect with disability, accommodation obligations or other human rights protections, the analysis becomes more demanding. Technological change does not displace those obligations.

The hardest cases may arise when AI makes an existing role unnecessary.

A company may legitimately eliminate a position because its remaining work can be distributed among fewer employees using AI. Employment law does not require an employer to preserve a position simply because it once had economic value, even if the incumbent is on maternity or another protected leave.

But again, an employer’s obligations do not disappear if the position is eliminated. Depending on the circumstances, the employee may be entitled to contractual or common law notice, statutory termination and severance entitlements, or other compensation.

There is a subtler problem which must be apprehended with employment law in mind. Sometimes the job has not disappeared. While its routine components have been automated, the remaining responsibilities now demand more judgment, different expertise or greater accountability.

That may call for retraining, reassignment or a redesigned role rather than straightforward elimination. Whether an employer must consider alternatives depends on the circumstances and applicable legal obligations. There is no universal requirement to create a new job whenever technology changes an existing one.

But a company that treats every AI-driven change as an automatic redundancy may overlook opportunities to retain valuable employees and misjudge its legal exposure.

The question is not simply whether AI can perform an employee’s former tasks. It is whether the employer has correctly identified the work that remains and the consequences of ending the existing relationship.

Employers may be tempted to solve this problem by inserting broader language into employment agreements, allowing management to change duties, introduce new technology and reorganize the workforce.

Careful drafting is helpful, but contractual language is no magic wand. Its effectiveness depends on its wording, the circumstances and the applicable law. A clause purporting to confer unlimited discretion on the employer, for example, will not achieve what its author intended.

AI adoption should therefore be treated as workforce redesign, with employment law considered before organizational decisions are implemented. The legal analysis must be built into the transformation.

Management should identify which tasks are being automated, which should be redesigned, establish whether roles are genuinely redundant and determine what notice, accommodation or other obligations may apply.

AI will allow some companies to accomplish more with fewer people. Others will use it to expand or develop new products and create work that was previously uneconomic. The effects will vary by industry and organization.

The strongest employers will redesign work deliberately, establish defensible performance standards, preserve valuable talent where appropriate, and understand the legal consequences before implementing change.

The objective is to build an organization capable of changing faster than its competitors without allowing the legal consequences of that change to undermine its gains.

The companies that understand this will be free to reinvent themselves. Those that do not may discover that while the technology saved them salary costs, the law presented them with a bill for how they did away with them.

Howard Levitt is senior partner of Levitt LLP, leading his teams of labour lawyers in Ontario, Alberta and British Columbia. Howard has appeared in more Supreme Court employment law cases and provincial appeals than any lawyer in Canadian history. A bestselling author, he discusses current workplace issues on the podcast, At Work with Howard Levitt.