‘Organizations struggle to see meaningful changes, and they struggle to act on the ones they see’
Artificial intelligence (AI) is changing jobs faster than organizations can govern them, leaving employees to absorb the risk and employers exposed to legal challenges, according to Ravinder Tulsiani, a Toronto-area learning and development leader.
“Where AI adoption is real, it is changing work faster than organizations can govern it, and the gap has two parts: organizations struggle to see meaningful changes, and they struggle to act on the ones they see."
Statistics Canada reported in June 2026 that 19.2 per cent of businesses used AI to produce goods or deliver services in the 12 months before its second-quarter survey, up from 12.2 per cent a year earlier. Yet only eight per cent of businesses told a December 2025 Bank of Canada survey of business leaders that they use AI significantly in their core operations.
As covered in our report on how job descriptions are years behind AI, outdated role documents are already slowing hiring.
Building a repeatable role audit
Tulsiani says frequency of AI use is the wrong test. “A workflow can change without materially altering a job.”
“The question is whether the change alters what someone is accountable for, the judgment expected of them, the capabilities the work requires, or the risks they manage,” he says.
On ownership, he says: “When a manager and HR disagree, that split decides it: the business defines what the work requires, HR defines what the formal role says.”
A trigger-based review assumes someone notices and reports the trigger, which fails exactly when employees adopt AI independently, Tulsiani says.
Organizations fall short in three predictable places, he says. “First, they audit the document against the document instead of the document against the work, comparing a job description to a template rather than to what people actually do.”
“Second, they run it as a project: one big rewrite, then nothing for two years. Third, they measure completion instead of accuracy: how many descriptions were updated, rather than whether the description matches the work,” he says.
Constructive dismissal can arise when an employer unilaterally makes a fundamental change to the terms of employment, and AI requirements can qualify, Aaron Zaltzman, an employment lawyer at Whitten Lubin in Toronto, told Canadian HR Reporter in coverage of Shopify’s AI mandate for staff.
Workers caught between old and new roles
When organizations fail to recognize material changes, employees carry the risk without the recognition, Tulsiani says.
He warns against cosmetic fixes. “A new title does not clarify responsibilities; a revised performance agreement, defined decision authority, and updated capability requirements do far more... What a material change in responsibilities does warrant is a compensation review.”
Asked whether HR models the agility it advocates, Tulsiani says: “No, not yet as a profession, though the picture varies by organization.”
“That structure explains part of the gap, but it does not absolve HR of responsibility for it,” he says, pointing to finance, technology and risk functions that shape budgets and AI policy.
“HR strengthens its credibility by applying the same standards internally,” he says.
