Nothing seems particularly different when you walk by the majority of corporate offices in Toronto’s financial district or midtown Manhattan. The structures continue to hum. Coffee is still prepared. People continue to gaze at screens and swipe badges.
However, something has changed within, somewhere between the quarterly reviews and the Slack messages. Autonomous AI agents have begun to perform tasks that were previously performed by humans in a quiet, methodical, and low-key manner. not helping. Not recommending. Acting.
| Category | Details |
|---|---|
| Topic | Autonomous AI Agents and Workforce Displacement |
| Year of Focus | 2026 — a defining turning point for AI adoption in enterprise settings |
| Key Companies Involved | Snowflake, Amazon, UPS, Target, General Motors, Goldman Sachs, Convictional, Anthropic |
| Estimated Job Risk (Goldman Sachs) | 6–7% of US workers face displacement due to AI adoption |
| Most Vulnerable Roles | Computer programmers, accountants, auditors, legal assistants, customer service reps |
| Biggest Disruption Zone | Middle management — not just entry-level technical roles |
| AI Adoption Reality Check | 56% of CEOs report no revenue or cost benefits from AI yet (PwC Global CEO Survey, 4,454 executives) |
| Notable Layoff Trigger | Amazon announced ~14,000 job cuts citing AI-driven efficiency in October 2024 |
| January 2025 Layoffs | Most mass layoffs in any January since 2009 in the US |
| Governance Status | Neither the US nor Canada has clearly defined legal rules for autonomous AI agents |
| Key Psychological Barrier | Worker trust — AI adoption fails when employees feel threatened or disempowered |
| Emerging Job Titles | Agent SRE, Chief Agent Officer |
| Companies Citing AI for Cuts | Pinterest, HP, UPS, Target, General Motors |
Today, Qaiser Habib, who oversees Canada engineering at Snowflake in Toronto, works with five AI agents for twenty to thirty hours a week. A year ago, his engineers continued to devote significant portions of their day to routine trend checks, dashboard scanning, and pursuing colleagues for data. The majority of that foundation has been transferred to machines.
These days, the typical engineer on his team collaborates with three or four agents every day to complete coding tasks under human supervision. “You don’t have to bother a human for basic questions anymore,” Habib replied. He describes it with a matter-of-factness that is neither triumphant nor troubled. Just made some adjustments.
Across industries, this shift is taking place, though not always with the same serenity. AI agents can plan, reason, adjust to changing business contexts, and finish multi-step tasks with little human oversight, in contrast to chatbots that wait for a prompt before responding.
They use meeting transcripts, internal databases, and calendars. These days, some of them oversee other agents; one creates code, another checks it for errors, and a third flags anything that requires a human signature. It is becoming more difficult to follow the chain of command.
It’s difficult to ignore the extra strain middle management is under. Some labor economists are directly attributing increased unemployment among recent graduates to automation, which is consistent with early predictions that AI would primarily eliminate entry-level technical jobs. However, Roger Kirkness, the founder of Convictional, a Toronto-based AI software company, contends that the more profound disruption is moving up the ladder.
The tools used by his company convert executive strategy into day-to-day operational tasks, which were previously solely the responsibility of supervisors, and provide assignments and feedback to staff members directly via an inbox interface.
“People are basically becoming managers of their prior jobs,” Kirkness stated. Humans are increasingly observing AI perform tasks and making decisions from a distance rather than writing code or creating assets. That is not insignificant. However, it is not the same.
Executives at Amazon cited AI’s ability to reduce layers and increase efficiency when they announced in October that the company planned to eliminate about 14,000 jobs. Similar announcements were made by UPS, Target, and General Motors. More mass layoffs occurred in January of this year than in any other January since 2009.
AI initiatives were mentioned as partial explanations by Pinterest and HP. According to Goldman Sachs, six to seven percent of American workers may eventually lose their jobs as a result of AI adoption; those most at risk are accountants, legal assistants, and customer service agents.
According to the bank, these effects could be “relatively temporary” as new roles become available. That is conceivable. For someone whose role has simply vanished, it’s also the kind of assurance that doesn’t mean much.
All of this has an odd tension to it. According to a PwC survey of over 4,000 executives in 95 countries, 56% of CEOs say they have not yet seen any quantifiable cost or revenue benefits from AI. Hiring decisions are being altered by the technology; they are being slowed down or eliminated, depending more on what businesses think AI will do in the future than on what it is doing now.
According to a Harvard Business Review survey published in December of last year, although AI hasn’t directly replaced workers, businesses have already started to reduce employment or halt hiring in anticipation of future capabilities. Losing a job due to the expectation of a machine rather than a machine itself is an odd form of displacement.
Behavioral scientist Stefano Puntoni of Wharton has been observing the effects of agent proliferation on workplace culture. According to his research, workers feel more at ease assigning tasks to AI than to their peers. No awkwardness, no social cost, and no fear of burdening someone. However, Puntoni is cautious about the meaning of that comfort. He contends that psychological issues are the true obstacle to adoption.
Workers’ sense of competence, autonomy, and belonging may be subtly threatened by generative AI. “If workers feel threatened, they may want the system to fail,” he stated. “At scale, that guarantees failure.” Kirkness takes that seriously enough that Convictional switched to a four-day workweek in order to share rather than extract productivity gains from AI. “Mass layoffs in the name of automation destroy trust,” he stated bluntly, as if he had witnessed it firsthand.
A different version of this tale is literally taking place on the streets, where delivery robots are trundling across college campuses and sidewalks in Los Angeles. There have been thousands deployed. Some become well-known for getting stuck, obstructing pedestrians, or annoying locals. Longtime Hollywood resident William Gude maintains a social media account called “Film The Robots LA” where he chronicles their mistakes.
The morning he spoke to a reporter, he got into a fight with one. However, the CEOs of these companies—Serve Robotics, Starship Technologies, and Coco Robotics—report an unexpected finding: the majority of people either assist or ignore the robots.
Starship claims that not a single unit has been stolen after nine million deliveries. It turns out that people are possibly nicer than they thought, and the robots are strangely charming.
It’s far less clear whether that kindness extends to the more general AI-and-work issue. Regarding autonomous agents, there is still no clear regulation in the US or Canada regarding what they can decide, who is responsible when they make mistakes, and how they should be governed. AI-driven insurance denials and hiring discrimination are already giving rise to legal challenges.
Depending on where you stand on the organizational chart, the fact that technology is advancing more quickly than the frameworks intended to contain it can either be exciting or concerning.
The majority of the experts keeping an eye on this field believe that companies that are positioning themselves well are the ones that treat governance as a basis for credibility rather than as a speed restriction. 2026 might be less forgiving than previous years for those who moved quickly and avoided responsibility.