Skip to content

Is AI the First Technology to Destroy More Jobs Than It Creates?

The Problem

Will AI create new jobs like every technology before it? Or are we seeing something different this time?

I’ve been thinking about this question a lot. The common wisdom says: “Don’t worry. Technology always creates more jobs than it destroys.” Cars replaced horses but created mechanics. The internet killed newspapers but created YouTubers.

But what if this time is different?

The Historical Pattern

Let me show you what happened with previous technology waves. Every time, the story was the same:

Historical Tech Revolution Pattern
┌─────────────────────────────────────────────────────────────┐
│ OLD TECHNOLOGY WAVE │
├─────────────────────────────────────────────────────────────┤
│ │
│ Technology Arrives │
│ │ │
│ ▼ │
│ Some Jobs Destroyed ──────────────┐ │
│ │ │ │
│ ▼ ▼ │
│ BUT... Technology Needs NEW JOBS! │
│ Human Help to Work: Created: │
│ - Maintenance - More than lost │
│ - Operation - New skill required │
│ - Infrastructure - New industries │
│ │
│ NET RESULT: More jobs than before │
└─────────────────────────────────────────────────────────────┘

Here’s a concrete comparison:

TechnologyJobs DestroyedJobs CreatedNet Effect
CarsCarriage drivers, stable handsDrivers, mechanics, gas station workers, road builders, auto factory workersPositive
InternetNewsstand workers, some retailWeb developers, delivery riders, streamers, content moderators, social media managersPositive
AIWriters, developers, analysts, support staff???Uncertain

See the pattern? Each technology needed humans to make it work. Cars needed drivers and mechanics. The internet needed developers and content creators.

Why AI Is Different

I think AI breaks this pattern in a fundamental way. Here’s the core difference:

AI Revolution Pattern
┌─────────────────────────────────────────────────────────────┐
│ AI TECHNOLOGY WAVE │
├─────────────────────────────────────────────────────────────┤
│ │
│ AI Arrives │
│ │ │
│ ▼ │
│ Jobs Destroyed ───────────────────┐ │
│ │ │ │
│ ▼ ▼ │
│ AND... AI's Core Feature: FEWER JOBS │
│ "Do More With Fewer People" Created: │
│ - Self-maintaining - Minimal │
│ - Self-improving - AI supervision │
│ - Scales without humans - Prompt engineers? │
│ │
│ NET RESULT: Fewer jobs than before? │
└─────────────────────────────────────────────────────────────┘

The key insight: AI’s value proposition IS reducing labor. A company buys AI tools specifically to do the same work with fewer people. This isn’t a bug—it’s the main feature.

Let me show you what I mean with real examples:

Traditional TechNew Human Roles RequiredAI EquivalentNew Human Roles?
CarDriver, mechanic, fuel attendantSelf-driving carMonitor? Maybe none
WebsiteDeveloper, designer, content writerAI-generated siteOne person instead of five
Customer serviceSupport team of 20AI chatbotOne supervisor instead of 20

The Self-Reinforcing Effect

Here’s what makes this even more concerning. AI doesn’t just fail to create jobs—it actively compresses existing ones.

I see this in my own work. Tools like Cursor and Claude Code multiply what one developer can do. A team of ten becomes a team of three. The work still gets done. But seven people are now looking for new roles.

The Compression Effect
Before AI:
Team A (10 developers) ────▶ Project completed
Time: 6 months
After AI:
Team B (3 developers + AI) ▶ Project completed
Time: 4 months
Quality: Same or better
Result: 7 fewer developer positions needed

This is happening across industries:

  • Content writing: One writer with AI does the work of five
  • Customer support: AI handles 80% of tickets, humans handle edge cases
  • Data analysis: AI generates reports that used to require analysts
  • Coding: AI assists developers, reducing team sizes

The One Possible Exception

I should mention one potential offset that some people discuss: “one-person companies.”

AI might enable more people to start businesses as individuals. Before, one person couldn’t build a company alone. Now, with AI tools, maybe they can.

One-Person Company Model
Traditional Startup:
Founder + 10 employees = Company
AI-Era Startup:
Founder + AI tools = Company

This could create self-employment opportunities. But I’m skeptical about the scale:

  • How many “one-person companies” can succeed?
  • Can this path absorb millions of displaced workers?
  • What happens when AI tools compete with AI tools?

No one has clear answers yet.

Why This Matters for You

Let me break down what this means for different groups.

For Workers

I think adaptability matters more than any specific skill now. The question isn’t “Will my job be automated?” but “How soon and how much?”

Consider this: your role might not be replaced by AI directly. It might be replaced by someone using AI who can do the work of three people.

For Investors

Returns may concentrate among AI enablers:

  • Chip makers (NVIDIA, AMD)
  • Compute providers (cloud services)
  • Foundation model companies

Application-layer startups face aggressive efficiency expectations. Investors want to see: “Do more with less.”

For Policymakers

The old playbook doesn’t work anymore. “Technology creates jobs” assumptions may not hold. We need to think about:

  • Safety nets for displaced workers
  • Retraining programs
  • Tax structures for a world with fewer jobs

Common Counterarguments

I hear three main objections to this view. Let me address each one.

”New jobs we can’t imagine will emerge”

This was true before. But AI’s rate of improvement is unprecedented. Any new job category could be automated before it reaches scale.

Think about it: “Prompt engineer” became a job. Now AI is getting better at prompting itself. The job appeared and might disappear within a few years.

”AI will augment humans, not replace them”

Augmentation is the transition phase. A human “augmented with AI” eventually becomes an AI system requiring minimal supervision.

The augmented human is a stepping stone, not an endpoint. Companies won’t pay for augmented humans when unaugmented AI gets the job done.

”Creative and emotional jobs are safe”

I wish this were true. But I see AI making progress on:

  • Writing (I’m using AI assistance right now)
  • Art and design
  • Music composition
  • Therapy and emotional support (AI companions)
  • Teaching and tutoring

No domain is guaranteed safe.

What I Think Will Happen

I don’t have certainty. But I see three possible futures:

Three Scenarios
┌─────────────────────────────────────────────────────────────┐
│ SCENARIO 1: History Repeats │
│ - AI creates new industries we can't imagine │
│ - Net employment stays stable or grows │
│ - Probability: Low (AI's nature is different) │
├─────────────────────────────────────────────────────────────┤
│ SCENARIO 2: Gradual Compression │
│ - Jobs decline slowly as AI improves │
│ - Society adapts with shorter work weeks, UBI │
│ - Painful transition but manageable │
│ - Probability: Medium │
├─────────────────────────────────────────────────────────────┤
│ SCENARIO 3: Rapid Displacement │
│ - AI capabilities accelerate │
│ - Millions displaced faster than new roles emerge │
│ - Significant social upheaval │
│ - Probability: Unknown │
└─────────────────────────────────────────────────────────────┘

Summary

In this post, I explored why AI might be the first technology in history that destroys more jobs than it creates. The key difference: AI’s core purpose is efficiency—doing more with fewer people. Unlike cars that needed drivers and mechanics, or the internet that needed developers, AI actively reduces human roles rather than creating them.

I think we’re entering genuinely new economic territory. The historical pattern of “technology creates jobs” might not apply this time. The one possible exception—one-person companies enabled by AI—remains unproven in scale.

What should you do? For workers: focus on adaptability. For investors: consider where value concentrates. For policymakers: prepare for scenarios where old assumptions fail.

The future isn’t certain. But I believe we should stop assuming AI will follow the pattern of every technology before it. The evidence suggests something different is happening.

Final Words + More Resources

My intention with this article was to help others share my knowledge and experience. If you want to contact me, you can contact by email: Email me

Here are also the most important links from this article along with some further resources that will help you in this scope:

Oh, and if you found these resources useful, don’t forget to support me by starring the repo on GitHub!

Comments