Part One of a Series on the Coming Labour Disruption
The Illusion of Corporate Loyalty: Why AI Will Be Different—and Worse
The Promise Nobody Should Believe
Something curious is happening in boardrooms, earnings calls, and corporate communications the world over. Companies that have spent decades demonstrating an almost clinical indifference to the fate of their workers are positioning themselves as guardians of the human workforce in the age of artificial intelligence. The language is warm, the pledges are generous, and the photography in the press releases is positively luminous. Reskilling. Upskilling. Human-AI collaboration. We are, apparently, all in this together.
The evidence suggests otherwise.
This is not cynicism for its own sake. It is a conclusion supported by more than a century of consistent corporate behaviour—behaviour that tells us, with considerable clarity, exactly what organizations do when technology makes workers redundant: they eliminate them, quickly, efficiently, and without sentiment. The current narrative of corporate loyalty to workers in the face of AI is not a moral evolution. It is a public relations strategy deployed at a moment of intense regulatory and reputational scrutiny. When that scrutiny fades—or when the competitive mathematics become too compelling to ignore—the pledge will dissolve as quietly as it was made.
What History Actually Shows
The idea that corporations have prioritized worker welfare over economic efficiency is largely a myth, and a recent one at that. The post-war decades in North America and Western Europe produced a brief period of relative stability in which large employers—particularly in manufacturing, utilities, and telecommunications—maintained something resembling long-term employment relationships. This was not altruism. It was the product of tight labour markets, strong unionization, and regulatory environments that made mass layoffs costly and politically difficult.
The moment those conditions changed, the behaviour changed.
By the early 1980s, the cultural transformation was complete. General Electric’s Jack Welch—admiringly nicknamed “Neutron Jack”—eliminated the jobs of one in four GE employees between 1980 and 1985, cutting 118,000 positions while the buildings remained standing (Uchitelle, 2006). He was not criticized. He was celebrated. Layoffs had been rebranded from a sign of failure to a signal of competitive seriousness. The doctrine spread rapidly across industries and decades.
The numbers tell the story with brutal efficiency. The U.S. economy averages approximately 9 million layoffs per quarter even in stable economic periods (Davis & von Wachter, 2011). During the recessions of 1990–91, 2001, and 2007–09, those figures spiked dramatically—the quarterly layoff rate rose by 208 basis points in the Great Recession alone (Davis & von Wachter, 2011). In 2009, employers initiated mass layoff events resulting in the separation of over 2.1 million workers in a single year (Bureau of Labor Statistics, 2013). IBM cut 60,000 employees in 1993 (Fortune, 2015). Citigroup shed tens of thousands during the financial crisis. General Motors, AT&T, Ford—the story repeats across every sector and every decade (Fortune, 2015).
The IBM case is particularly instructive because IBM had, for decades, maintained an informal policy of lifetime employment (Marketplace, 2016). Workers built careers there. Families built their lives around the assumption of stability. Then in 1993, the company cut 60,000 workers in a single stroke (Marketplace, 2016). The loyalty that employees had extended over lifetimes was not reciprocated. It never is when the balance sheet demands otherwise.
As MIT Sloan Management Review documented as far back as 1997, the era of corporate downsizing produced a fundamental and lasting collapse in employee loyalty—not because workers became ungrateful, but because they were responding rationally to the reality that their employers had abandoned the implicit contract first (Reichheld, 1997).
The New Loyalty Narrative—and Why It Exists
Against this historical backdrop, the current wave of corporate pledges to protect and retrain workers in the age of AI deserves serious scrutiny.
Amazon’s Upskilling 2025 program is the most frequently cited example of corporate responsibility in the AI era. Launched in 2019, it pledged $700 million to retrain 100,000 U.S. workers—roughly one-third of its American workforce at the time (Wharton School, 2025). Amazon later reported exceeding the target, with over 700,000 employees trained globally (Amazon, 2025). In 2025, a new initiative—Future Ready 2030—was announced, pledging $2.5 billion to upskill 50 million people globally (Allwork.Space, 2025).
The numbers sound significant until examined carefully. The original $700 million program worked out to approximately $1,200 per employee per year (Wharton School, 2025)—roughly the cost of a single community college course. Meanwhile, Amazon has simultaneously been one of the most aggressive deployers of warehouse automation in the world, systematically replacing the very roles those workers hold. Only 20% of workers displaced by automation have the educational background required to transition to high-tech roles (Nagar, 2025). McKinsey estimates that 70% of new AI-created jobs will require a college degree (Nagar, 2025). The arithmetic is not close, and it was never designed to be.
The more uncomfortable explanation for the loyalty narrative is not moral awakening but strategic positioning. Organizations are operating under significant scrutiny from regulators, governments, and the public. The EU AI Act, executive orders in the United States, and mounting political pressure around AI displacement have made optics genuinely costly. A company visibly and aggressively cutting its workforce to install AI systems faces reputational and regulatory risk that a company publicly committed to “human-AI collaboration” does not. The loyalty narrative is, at its core, a risk management strategy.
Anticipatory Cuts: The Eagerness Behind the Curtain
If the loyalty narrative were sincere, one would expect organizations to delay workforce reductions until AI systems had demonstrably replaced the relevant functions. The evidence suggests the opposite is occurring—and with an almost eager anticipation that reveals where organizational priorities actually lie.
Harvard Business Review’s analysis of a survey of 1,006 global executives found that companies are announcing layoffs or freezing recruitment not because AI is currently replacing jobs, but in anticipation of what it will eventually be able to do (Harvard Business Review, 2026). The cuts are getting ahead of the technology, not following it. As one analysis put it bluntly: leading CEOs—including those from Ford, Amazon, Salesforce, and JP Morgan Chase—have proclaimed that many white-collar jobs at their companies will soon disappear, and are acting accordingly (Harvard Business Review, 2026).
The scale is already substantial. In 2025 alone, 54,836 AI-linked layoffs were announced, with a further 502,000 planned for the following year (LinkedIn, 2025). Amazon confirmed 16,000 corporate job cuts in January 2026, leaving open the possibility of further reductions as it pursues AI-driven transformation (Reuters, 2026). Autodesk cut approximately 1,000 jobs to reallocate funding to AI projects. Dow cut 4,500 jobs—13% of its entire workforce—citing AI and automation. Snap eliminated 16% of its global workforce, with its CEO explicitly citing rapid AI advancements as the rationale (Business Insider, 2026). Meta cut over 1,000 jobs from its Reality Labs division as it shifted focus from the metaverse to AI (Reuters, 2026).
What makes these cuts particularly telling is not their scale but their timing. The AI models being used are, by the companies’ own acknowledgment, not yet ready to perform the work being eliminated. Workers are being shown the door while their replacements are still being built. This is not cautious transition management. It is eager anticipation—organizations positioning themselves to extract the efficiency gains of AI the moment capability crosses the threshold, having already cleared the human inventory in advance.
The stock market has not been an innocent bystander to this process. When Jack Dorsey announced that his company Block would lay off 40% of its staff due to artificial intelligence, the firm’s stock price skyrocketed—an event experts described as a potential “tipping point” signalling to every other board of directors the reward awaiting similar announcements (Global News, 2026). Investor enthusiasm for companies that implement AI-linked workforce reductions has been widely documented: numerous instances have been observed where announcement of AI integration coincided with rising stock prices (Forbes, 2025). The financial markets are, in effect, offering a bounty for workforce elimination—and organizations are collecting it.
A bandwagon effect has emerged in parallel. Research from the Wharton School found that companies observe competitors cutting jobs and feel compelled to follow suit regardless of their own operational readiness, with investors interpreting layoff announcements as signals of proactive efficiency improvement (CNBC, 2025). The result is an accelerating cycle of competitive signalling in which workforce reduction becomes a performance of AI seriousness as much as a genuine operational response to technological change.
Why the Competitive Mathematics Make Loyalty Impossible
Even granting the most charitable interpretation—that some organizations genuinely intend to honour their loyalty pledges—the competitive structure of AI adoption makes those pledges structurally impossible to keep. This is perhaps the most important and least discussed dimension of the coming displacement.
When AI makes a role economically viable to automate, no individual organization can afford to delay indefinitely. The mechanism works as follows: a company that automates reduces its labour costs; under competitive pricing, those cost savings are passed on to consumers through lower prices, winning market share; the competitor that has not automated loses share and faces a choice between automating or contracting. The competitor that chooses loyalty to its workforce is gradually competed out of the market by the competitor that does not. No board of directors can justify sustained higher labour costs to shareholders when the alternative exists next door—not because boards are uniquely callous, but because the fiduciary structure of publicly traded companies makes that justification legally and practically untenable.
A 2026 economics paper formalized this dynamic with particular precision, showing that AI-driven automation creates what economists term a demand externality—a situation in which one company’s automation decision imposes costs on other companies that the first company does not have to pay (University Herald, 2026). When a firm replaces workers with AI, the displaced workers are also consumers who stop buying things. That reduction in consumer spending harms every company in the sector, including the automating firm—but the automating firm bears only a fraction of that harm while capturing the full benefit of its cost savings (University Herald, 2026). Every firm faces this same calculation. Every firm therefore automates beyond what would be collectively optimal—and, crucially, they know they are doing it.
The paper’s authors call this the Red Queen effect, after the Lewis Carroll character who must keep running faster simply to stay in the same place (University Herald, 2026). More competition between firms makes the problem worse, not better. A monopoly firm actually has an incentive to avoid over-automation because it fully internalizes the demand destruction it creates—it is hurting its own customers. The more fragmented the market, the more the demand destruction is externalized onto competitors, and the wider the gap between what individual firms choose to automate and what would be best for the economy as a whole (University Herald, 2026). The loyalty promise is not just hollow—it is mathematically impossible to honour in a competitive market once AI capability crosses the replacement threshold. The organization that tries will not survive to keep the promise.
The Union Question—and Why It Has No Good Answer
For workers who look to organized labour as the traditional bulwark against this kind of displacement, the picture is grimmer than most commentators acknowledge.
Union power historically rested on a specific and now-fragile assumption: that the skills workers held were difficult to replace quickly and that disruption was costly to the employer. Strikes worked because they imposed genuine economic pain. The implicit threat—”you need us more than you can admit”—had real force when that was true.
AI dismantles that assumption at its foundation. The moment a role can be automated, the union’s leverage in that role evaporates. Worse, AI enables something qualitatively new: a competitor can enter a unionized sector by building an AI-native operation from scratch, bypassing the existing workforce entirely. The union does not lose a negotiation. It loses the industry.
This is not hypothetical. In 2023, the National Eating Disorders Association replaced its entire helpline workforce with an AI chatbot—four days after those workers won their union election (UnionTrack, 2023). The message was unambiguous: the existence of the union triggered immediate replacement, not negotiation. Amazon has deployed AI-powered surveillance systems to monitor worker communications and predict union organizing behaviour before it becomes visible through conventional means (Power at Work, 2025). AI is enabling what analysts have described as a twenty-first-century variant of classic union suppression (UnionTrack, 2023).
The broader pattern is predictable: any sector facing union challenges that can be disrupted by an AI-native competitor will be. A new entrant does not inherit the labour relations of the incumbent. It inherits only the cost advantages of having none. The union is not fighting the employer. It is fighting the employer’s ability to stop needing it.
Why the Public Accepts the Narrative
If the evidence against corporate loyalty is this abundant and this consistent, why does the reassurance narrative land so effectively? Why do workers—intelligent, experienced, and increasingly informed workers—accept the promise that their employers will protect them through the AI transition?
The answer lies not in credulity but in psychology. Optimism bias—the well-documented human tendency to believe that bad outcomes are systematically more likely to happen to others than to oneself—is one of the most robust findings in behavioural science (Sharot, 2011). It applies to health outcomes, financial risks, and, critically, to AI job displacement. “It will be someone else’s role that gets automated” is not an unusual thought. It is the default cognitive position of almost every worker confronting this question.
Paired with optimism bias is the information avoidance instinct—what psychologists call the ostrich effect. People deliberately avoid or downplay negative information when knowing it would be distressing or force uncomfortable behavioural change (Golman et al., 2017). Corporate loyalty narratives are, from this perspective, a gift to the avoidance instinct. They provide a socially sanctioned reason not to engage with threatening information: “My company is handling this. I don’t need to worry.”
A further mechanism compounds both of these: the tendency to conflate intent with outcome. Organizations that genuinely intend to protect their workers may do so sincerely. But as the competitive dynamics described above make clear, sincere intent and structural possibility are different things. A company can mean every word of its reskilling pledge and still be forced by competitive mathematics to abandon it. Workers who accept the narrative based on perceived sincerity are misreading the situation—not because they are foolish, but because they are applying a human relational framework to a structural economic problem that operates by different rules.
The Conclusion the Evidence Demands
No institution is coming to save the individual worker from AI displacement. Not their employer, whose loyalty has always been conditional on economic convenience and will be again. Not the union, whose leverage depends on a scarcity of skills that AI is systematically eliminating. Not the government, whose policy responses lag technological reality by years and whose retraining programs produce, at best, a 30% success rate in comparable employment outcomes (MIT Task Force on the Work of the Future, 2020).
The only protection that travels with a worker from role to role, employer to employer, and industry to industry is the quality of their own thinking—the capacity to analyze, to frame problems, to adapt, to create, and to ask questions that have not yet been asked. These are not soft skills. The World Economic Forum has ranked analytical thinking, creative thinking, resilience, and curiosity as the most sought-after capabilities by employers year after year (World Economic Forum, 2025). They are precisely the capabilities that AI cannot yet replicate at a meaningful human level, and there is genuine reason to believe the window for developing them—before the displacement curve steepens—is narrower than most people currently assume.
The loyalty organizations are performing right now is a production. The curtain will come down when the economics demand it. The question every worker should be asking is not whether their employer means what it says. It is whether they have made themselves into something no employer—and no algorithm—can afford to lose.
References
Allwork.Space. (2025, October 23). Amazon launches $2.5B Future Ready 2030 program to upskill 50 million people amid automation push. https://allwork.space/2025/10/amazon-launches-2-5b-future-ready-2030-program-to-upskill-50-million-people-amid-automation-push/
Amazon. (2025, June 30). Amazon has upskilled over 700,000 employees globally through Upskilling 2025. https://www.aboutamazon.com/news/workplace/amazon-employees-upskilling-education-training
Bureau of Labor Statistics. (2013, January 4). Reasons for extended mass layoffs, 2009–2011. U.S. Department of Labor. https://www.bls.gov/opub/ted/2013/ted_20130104.htm
Business Insider. (2026, April 19). 10 companies that have said they’re doing AI-related layoffs. https://www.businessinsider.com/list-companies-replacing-human-employees-with-ai-layoffs-workforce-reductions
CNBC. (2025, November 4). AI-washing and the massive layoffs hitting the economy. https://www.cnbc.com/2025/11/04/white-collar-layoffs-ai-cost-cutting-tariffs.html
Davis, S. J., & von Wachter, T. (2011). Recessions and the costs of job loss. Brookings Papers on Economic Activity, 2011(2), 1–72. https://www.brookings.edu/wp-content/uploads/2011/09/2011b_bpea_davis.pdf
Forbes. (2025, May 4). Companies are replacing workers with AI—fast. https://www.forbes.com/sites/jackkelly/2025/05/04/its-time-to-get-concerned-klarna-ups-duolingo-cisco-and-many-other-companies
Fortune. (2015, September 20). The 10 biggest corporate layoffs of the past two decades. https://fortune.com/2015/09/20/biggest-corporate-layoffs/
Global News. (2026, February 27). Could Block’s AI-fueled layoffs be a ‘tipping point’ for white-collar jobs? https://globalnews.ca/news/11710322/ai-layoffs-future-experts/
Golman, R., Hagmann, D., & Loewenstein, G. (2017). Information avoidance. Journal of Economic Literature, 55(1), 96–135. https://doi.org/10.1257/jel.20151245
Harvard Business Review. (2026, February 2). Companies are laying off workers because of AI’s potential—not its performance. https://hbr.org/2026/01/companies-are-laying-off-workers-because-of-ais-potential-not-its-performance
Marketplace. (2016, June 13). IBM: When corporations took care of their employees. https://www.marketplace.org/story/2016/06/13/profit-ibm
MIT Task Force on the Work of the Future. (2020). The work of the future: Building better jobs in an age of intelligent machines. Massachusetts Institute of Technology.
Nagar, S. (2025, January 9). AI & the workforce—the illusion of reskilling & the coming crisis. https://sumirnagar.com/2025/01/10/ai-the-workforce-the-illusion-of-reskilling-the-coming-crisis/
Power at Work. (2025, October 25). Big tech wants AI without rules. Here’s how workers are fighting back. https://poweratwork.us/big-tech-wants-ai-without-rules-heres-how-workers-are-fighting-back
Reichheld, F. F. (1997). Loyalty in the age of downsizing. MIT Sloan Management Review, 38(4), 29–37. https://sloanreview.mit.edu/article/loyalty-in-the-age-of-downsizing/
Reuters. (2026, February 25). Factbox—Companies cutting jobs as investments shift toward AI. https://ca.finance.yahoo.com/news/factbox-companies-cutting-jobs-investments-163836089.html
Sharot, T. (2011). The optimism bias. Current Biology, 21(23), R941–R945. https://doi.org/10.1016/j.cub.2011.10.030
Strategic Market Research. (2026). AI replacing jobs statistics 2025: 300 million jobs at risk. https://www.strategicmarketresearch.com/blogs/ai-job-loss-statistics
Uchitelle, L. (2006). The disposable American: Layoffs and their consequences. Knopf.
UnionTrack. (2023, December 4). Artificial intelligence in union busting: A new tool for employers. https://uniontrack.com/blog/union-busting
University Herald. (2026, April 18). Why rational firms keep automating even when it destroys demand. https://www.universityherald.com/articles/80291/20260419/why-rational-firms-keep-automating-even-when-it-destroys-demand-new-eco
Wharton School. (2025, October 27). Will Amazon’s plan to ‘upskill’ its employees pay off? University of Pennsylvania. https://knowledge.wharton.upenn.edu/podcast/knowledge-at-wharton-podcast/amazon-retraining-employees/
World Economic Forum. (2025). The future of jobs report 2025. https://www.weforum.org/publications/the-future-of-jobs-report-2025/