AI Replacing Jobs in Canada? Follow the Money — Who Benefits When You’re Afraid

Published: April 2026 | By: Zeus eBikes Canada

You have been told to be afraid of artificial intelligence.

You have been told it will take your job. Destroy your industry. Maybe end civilisation. The headlines are everywhere — on the front page of newspapers owned by billionaires who are, at this very moment, investing billions into the same technology they are telling you to fear.

We are an eBike company. We are not AI researchers. We are not policy analysts. We are two brothers in Canada who use AI every day to run our business, and who noticed something that did not add up: the people with the most money and power are the loudest voices telling ordinary Canadians to be afraid of the one technology that could make money and power less relevant.

So we did what we do with every guide we publish — we followed the evidence. We traced the money. We checked who funds the fear, who profits from it, and what the actual data says about AI and Canadian jobs.

We are not going to tell you what to think. We are going to show you what we found. You decide what it means.

1. The Contradiction: Fear in Public, Billions in Private

The people telling you AI is dangerous are the same people spending more money on AI than on anything else. This is not an interpretation. It is a matter of public record.

Elon Musk

In March 2023, Musk co-signed a public letter from the Future of Life Institute calling for a six-month pause on AI development more powerful than GPT-4. The letter stated: “AI systems with human-competitive intelligence can pose profound risks to society and humanity.”

Four months later — July 2023 — Musk founded xAI, a new artificial intelligence company. He hired engineers from Google DeepMind. He built a 100,000-GPU supercomputer in Memphis. By 2024, xAI had raised over $6 billion in venture funding. By late 2024, it was valued at $50 billion.

He signed a letter asking everyone to pause. Then he built a competitor as fast as possible.

Jeff Bezos

Bezos owns the Washington Post, which publishes AI fear stories regularly — job displacement, bias, deepfakes, civilisational risk. Standard news coverage.

Bezos also invested $8 billion in Anthropic (an AI company) through Amazon — $4 billion initially in 2023, doubled to $8 billion by November 2024. Amazon Web Services — the cloud infrastructure that most AI companies run on — generates approximately $100 billion per year in revenue. Amazon has deployed 750,000+ robots in its warehouses, replacing picker and packer jobs at a scale that dwarfs anything ChatGPT has done to white-collar employment.

The same person owns a newspaper that runs AI fear stories and holds $8 billion in AI investments through Amazon. The Post maintains an editorial independence policy — but the structural conflict of interest is a matter of public record.

Sam Altman — OpenAI

In May 2023, Altman testified before the U.S. Senate and said: “If this technology goes wrong, it can go quite wrong.” He proposed a federal licensing agency for AI. He called for mandatory safety audits.

In the twelve months following that testimony, he launched GPT-4 Turbo, GPT-4 Vision, DALL-E 3, custom GPTs, expanded API access, signed enterprise deals worth hundreds of millions, and raised additional billions. OpenAI’s valuation reached $157 billion.

The licensing regime he proposed would require testing infrastructure, legal compliance teams, and safety audits that cost millions — trivial for OpenAI (backed by $13 billion from Microsoft), fatal for a startup in a garage.

7:20 PM. Bay Street. The tower has 30 floors of people writing the rules. He is on the sidewalk reading them. · Playcut.ai

2. The Pause Letter Nobody Honoured

On March 28, 2023, the Future of Life Institute published an open letter calling for a six-month pause on AI training. It was signed by over 30,000 people, including Elon Musk, Steve Wozniak, and Yoshua Bengio.

Here is what the signatories did during those six months:

• Musk — founded xAI and began recruiting AI engineers
• Emad Mostaque (then-CEO of Stability AI) — continued training and releasing image generation models without pausing
• Stuart Russell (UC Berkeley) — continued publishing AI research papers
• Yoshua Bengio — continued AI research and co-authored papers on advanced AI systems

Who did not sign: Sam Altman (OpenAI), Satya Nadella (Microsoft), Sundar Pichai (Google), Mark Zuckerberg (Meta) — the people actually building the most powerful systems. They had no incentive to pause. They were ahead.

Who funded the organisation that wrote the letter: The Future of Life Institute received $10 million from Elon Musk in 2015. Its other major funder is Jaan Tallinn, co-founder of Skype, who is also an investor in AI companies including Anthropic.

The letter asked for a pause. Nobody paused. The letter’s biggest funder started a competing AI company. Make of that what you will.

3. Regulation as a Weapon Against the Little Guy

There is a word in economics for what happens when an industry helps design the rules that govern it: regulatory capture. It is not a conspiracy theory. It is a documented phenomenon studied for decades. And it is happening in AI right now.

The playbook is simple

1. Call for regulation (to appear responsible)
2. Propose compliance requirements that cost millions (safety audits, licensing, documentation)
3. Absorb those costs easily because you are a billion-dollar company
4. Watch smaller competitors collapse under the same requirements

The numbers

Source: OpenSecrets, 2023 federal lobbying data. All companies lobbied on multiple tech issues; AI was a primary focus.

The table where $14.5 million in lobbying becomes language. The coffee ring on the mahogany belongs to someone comfortable enough not to use a coaster. · Playcut.ai

The EU AI Act — the world’s most comprehensive AI regulation — imposes compliance costs estimated at €400,000 per high-risk AI system (Centre for Data Innovation, 2021). Google — which spent $14.5 million on lobbying in 2023 alone — can absorb this. A two-person company in Saskatoon cannot. The same rules apply to both.

This is not unique to AI. It has happened before:

• Banking: Dodd-Frank (2010) imposed compliance costs of $10,000–$50,000 per employee per year for small banks versus $1,000–$5,000 for large banks (economies of scale). JPMorgan CEO Jamie Dimon said it openly: “In one way, [regulation] creates a bigger moat.”
• Pharmaceuticals: FDA compliance costs favour large pharma. “Pay-for-delay” agreements between brand-name and generic companies cost U.S. consumers an estimated $3.5 billion per year in higher drug prices (FTC).
• Taxis: New York taxi medallions cost over $1 million — defended as “public safety” until ride-sharing proved the barrier was about money, not safety.
• Telecom: Municipal broadband has been banned or restricted in at least 16 U.S. states through laws lobbied for by incumbent ISPs.

When the biggest company in an industry asks to be regulated, examine who the regulation eliminates.

2 AM. One monitor. One person. The $100/month tool doing what used to require a $10 million operation. Nobody knows this room exists. That is the point. · Playcut.ai

4. The Media Machine: Who Profits From AI Fear

AI fear is not just a story. It is a business model.

The ownership map

These outlets run AI fear headlines. Their owners invest billions in AI. We are not alleging that any owner directs editorial coverage to serve their investments. Most maintain editorial independence policies, and we take those policies at face value. But the structural conflict of interest exists — the same corporate families that profit from AI also own the outlets shaping public perception of AI — and you are entitled to know about it.

Why fear sells

This is not conspiracy. It is economics.

• Fear headlines generate significantly more engagement than balanced coverage — research consistently shows negativity bias drives 2–5× higher click-through rates (documented across multiple digital news studies, including findings cited by the Reuters Institute)
• False or fear-based content spreads 6× faster on social media than factual content (MIT, published in Science, Vosoughi et al. 2018)
• Ad-supported media profits from engagement. Fear drives engagement. Fear drives revenue.

“AI will probably change your job somewhat over the next decade” does not get shared. “AI will take your job” gets 500,000 clicks. The incentive structure is plain.

The Goldman Sachs headline

In March 2023, Goldman Sachs published a report that was covered by every major outlet with the headline: “AI could replace 300 million jobs.”

What the actual report said:

• AI could “expose” 300 million jobs to automation — meaning some tasks within those jobs could change. Not that 300 million people would be fired.
• The same report estimated AI would increase global GDP by 7% (~$7 trillion).
• The same report noted that historically, technology that displaces jobs creates more new jobs than it destroys — citing ATMs (which increased total bank employment) and computerisation (which created the entire IT industry).

The nuance was stripped. The fear number led every headline. The GDP growth and job creation findings were buried.

6:15 AM. Yonge Street. Thirty headlines designed to stop him. He kept walking. · Playcut.ai

5. Canadian Media — Follow the Ownership

Since this article is for Canadians, here is what we found about Canadian outlets specifically.

• The Globe and Mail — owned by the Woodbridge Company (Thomson family). The Thomson family controls Thomson Reuters, which has invested hundreds of millions in AI-enhanced legal and financial products including CoCounsel (an AI legal research tool built on GPT-4, acquired for $650 million in 2023). The Globe runs AI disruption stories. The family profits from AI integration. The Globe and Mail maintains editorial independence from its ownership’s other business interests — the conflict we are noting is structural, not editorial.
• CBC — publicly funded. No direct AI investment. However, CBC as a media institution is threatened by AI-generated content and declining ad revenue, creating an institutional bias toward framing AI as a threat to media. CBC has explored AI tools internally for transcription and content production while covering AI as disruptive.
• Postmedia (National Post, etc.) — major shareholder is Chatham Asset Management, a U.S. hedge fund. Postmedia has been financially distressed for years. AI-generated content threatens their already fragile business model. This institutional anxiety is visible in coverage.

This does not mean Canadian media coverage is wrong. It means Canadian media institutions have financial and institutional reasons to emphasise AI fear over AI opportunity. You are entitled to know that context when reading their coverage.

6. The Real Threat: AI Removes Gatekeepers — And the Gatekeepers Know It

Here is what the fear campaign does not want you to see. AI is not primarily replacing workers. It is removing intermediaries who charged rent on access to knowledge and capability.

The corridor was never designed for him. The door at the end was never left open for him. He walked through it anyway. · Playcut.ai

Legal industry

A top-tier law firm partner bills $1,500–$2,500 per hour. Kirkland & Ellis, the world’s highest-grossing law firm, reported $7.2 billion in revenue in 2023. Average partner compensation exceeded $7 million.

AI legal tools can now review contracts, draft memos, and conduct legal research in minutes. What cost $100,000–$500,000 for M&A due diligence can be done for a fraction of that cost.

The response: State bar associations threatened DoNotPay’s founder with jail time for building an AI that contested parking tickets — not because the AI was wrong, but because it bypassed a lawyer intermediary. The American Bar Association spent $850,000 on federal lobbying in 2023 (OpenSecrets). The argument was never about quality. It was about preserving the intermediary.

The Legal Services Corporation’s 2022 Justice Gap Study found that 92% of low-income Americans’ civil legal problems received inadequate or no legal help. AI could close this gap. But it would also collapse the fee structure that generates $450 billion per year in U.S. legal services revenue.

Healthcare

An AI system called IDx-DR received FDA clearance in 2018 as the first autonomous AI diagnostic system — it screens for diabetic retinopathy from retinal images at approximately $39 per screening. A specialist ophthalmologist charges $100–$300 for the same screening. Google DeepMind’s AI matched or exceeded world-leading ophthalmologists in diagnosing over 50 eye diseases. Stanford’s dermatology AI matched 21 board-certified dermatologists in identifying skin cancer.

The response: The American Medical Association spent $21.2 million on federal lobbying in 2023 (OpenSecrets) — one of the top-spending lobbying organisations in the U.S. Their consistent position: AI must operate “under physician supervision.” This preserves the physician as a mandatory intermediary. The average radiologist earns $530,000 per year (Bureau of Labor Statistics, 2023). The WHO estimates a global shortage of 10 million health workers by 2030 (revised from an earlier 18 million projection). In rural Canada, many communities have no dermatologist within 100 kilometres.

AI could serve them. But it cannot legally operate without a physician in the loop — and the physician is not available.

Consulting

McKinsey charges $500,000–$4 million for a strategy engagement. In September 2023, BCG published its own study (with Harvard and Wharton researchers) showing that consultants using GPT-4 completed tasks 25% faster and produced 40% higher-quality results.

BCG’s response: build proprietary AI tools internally to capture the value — rather than let clients use AI directly and skip the $500K engagement. Accenture spent $2.7 million on U.S. federal lobbying in 2023 while positioning itself as the company you need to hire to implement AI. The consulting industry’s message: “AI is complex, you need us.” The reality: AI is the tool that lets you skip them.

Education

A Harvard MBA costs approximately $230,000. Khan Academy’s AI tutor (built on GPT-4) provides personalised, one-on-one tutoring for $4 per month. A private human tutor costs $40–$150 per hour.

Multiple universities banned AI outright (Sciences Po in January 2023, University of Hong Kong in February 2023). Australian universities returned to handwritten, in-person exams. The credential — the degree — is a gatekeeping mechanism. When a student in Sudbury can learn the same material as a student at Harvard for 0.002% of the cost, the $230,000 price tag is not for education. It is for the credential — a social sorting mechanism that AI threatens to expose.

Left: $230,000 to sit here. Nobody is sitting. Right: $4 a month. Someone is learning. · Playcut.ai

The pattern across every industry

Not a single major industry incumbent has argued that AI should be banned because it produces worse results. The arguments are always about “safety,” “oversight,” and “supervision” — all of which require the incumbent to remain in the loop and continue collecting fees.

7. The Stories Nobody Covers

For every AI fear headline, there is a story about someone whose life improved that did not get published. Fear clicks. Empowerment does not. Here are the receipts.

Be My Eyes — 500,000 blind users

Be My Eyes, a Danish app, integrated GPT-4’s vision capabilities in March 2023. Over 500,000 blind and low-vision users can now point their phone camera at a menu, a medicine bottle, or a street sign and receive instant audio description — without waiting for a human volunteer. Previously, this required either a sighted companion or a wait for an available volunteer. AI did not take a job. It gave half a million people independence.

High Park. Cherry blossom season. She pointed her phone at something she has never seen. The AI described it. He watched her face when she heard it. · Playcut.ai

Niramai — breast cancer screening without a clinic

Niramai, founded in India by Geetha Manjunath, uses AI and thermal imaging to screen for breast cancer. No radiologist required. No $500,000 mammography machine. Works in rural clinics. Screened over 400,000 women as of 2025. Detects tumours up to five years earlier than physical examination. In regions where the nearest radiologist is a three-hour drive, this is not a convenience. It is the difference between detection and death.

A $200 device. A trained woman. A life potentially saved. No radiologist in the room. No $500,000 machine. This is what access looks like when the gatekeepers are not invited. · Playcut.ai

AlphaFold — 200 million protein structures, free

Google DeepMind’s AlphaFold solved the protein folding problem in 2020 and released 200 million protein structure predictions for free. Researchers in 190 countries who could never afford X-ray crystallography lab equipment now have access to the same structural data as Harvard and MIT. Over half a million researchers accessed the database in its first year, surpassing two million within two years. This would have been impossible without AI. It cost researchers nothing.

One-person businesses reaching $1M+ revenue

Danny Postma (Netherlands) built HeadshotPro — AI professional headshots — as a solo operator. Revenue reportedly reached $1 million within months. One person. No employees. No venture capital. Competing against photography studios that charge $200–$500 per headshot session. Pieter Levels, another solo developer, publicly crossed $200,000 per month in revenue in 2024 across multiple AI-powered products — with zero employees.

Khan Academy — a $4/month tutor for every student

Khan Academy deployed an AI tutor (Khanmigo, built on GPT-4) for 40,000+ students. Cost: $4 per month. A private human tutor costs $40–$150 per hour. Sal Khan explicitly frames this as giving every child the 1-on-1 tutoring that was previously available only to families who could afford it. The equaliser the education industry never built — because building it would have undercut their pricing.

8. What the Canadian Data Actually Says

You have heard the fear. Here is the data.

Employment

• Statistics Canada Labour Force Survey (2023–2025): Canada’s unemployment rate fluctuated between 5.0% and 6.9% — within normal cyclical range. No StatCan report has attributed measurable job losses specifically to AI deployment.
• U.S. Bureau of Labor Statistics: The BLS Occupational Outlook Handbook (2024) projects growth in most white-collar fields through 2032, including many that AI fear stories say will be eliminated. The “AI job apocalypse” narrative has not appeared in any BLS statistical release.
• McKinsey Global Institute (June 2023): Estimated 12 million occupational transitions by 2030 — comparable to the rate of change during COVID and previous economic shifts. Projected AI would create new job categories, not just destroy existing ones. The media headlined the transition number and buried the creation finding.

AI adoption

• Statistics Canada (Q2 2025): 12.2% of Canadian businesses used AI — double the 6.1% from Q2 2024. Top applications: text analytics (35.7%), data analytics (26.4%), chatbots (24.8%), marketing automation (23.1%). Only 6% of AI-adopting businesses reported actual layoffs.
• 57% of Canadian employees see AI as positive. 54% are comfortable with AI collaboration (Abacus Data).

The fear gap

Pew Research Center (2024): 52% of Americans are “more concerned than excited” about AI. But among people who actually use AI tools regularly, that number drops to 28%. Direct experience reduces fear. Media consumption increases it.

Read that again. The people who use AI are less afraid. The people who read about AI are more afraid. The gap between the two groups tells you where the fear is coming from.

You Decide

We did not write this article to tell you AI is safe. We did not write it to tell you AI is dangerous. We wrote it because Canadians deserve to see the full picture — the money, the contradictions, the data, and the stories that fear-based coverage leaves out.

Here is what we know:

• The people telling you to fear AI are investing more money into it than any technology in human history.
• Not one person who signed the AI pause letter paused anything.
• The regulations being proposed by AI company executives would benefit their companies and burden their competitors.
• AI fear headlines get 2–5× more clicks than balanced coverage. Media runs on clicks.
• Canadian unemployment is within normal range. AI-adopting businesses report 6% layoffs, not the apocalypse.
• People who use AI are less afraid than people who read about it.
• AI’s biggest documented impact is not replacing workers — it is removing gatekeepers who charged rent on access to knowledge.

Make of that what you will. We are not here to tell you what to think. We are here to make sure you have what you need to think for yourself.

7:50 AM. Tuesday. April. A man on a bicycle, going to work. Not afraid. Not performing courage. Just using the tool. The boring truth nobody publishes. · Playcut.ai

Every claim in this article is sourced from public records, published reports, SEC filings, Congressional testimony, peer-reviewed research, or official statistical releases. Where we infer motivation, we identify the inference. Where evidence is incomplete, we say so. We are not AI researchers, policy analysts, or economists. We are two brothers who followed the money and wrote down what we found. If we got something wrong, call us at 1-866-938-7580 and we will correct it.