Published
Topics
AI jobs future of work career transitions AI unemployment job displacement reskilling

The Messy Middle: AI's Impact on Jobs (2025-2035)

History says you'll have a job in 2035. It just might not be the one you trained for. Here's what the transition period looks like—and how to navigate it.

"Will AI take my job?"

It's the question on everyone's mind. The answer is both reassuring and
unsettling: History says yes, jobs will exist. But the transition period will
be messy, uneven, and challenging.

Let me show you what the data tells us—and what it doesn't.

The Scale of Disruption

Estimates suggest that globally, around 300 million full-time jobs could be
affected by generative AI automation in the coming years.

"Affected" doesn't mean "eliminated"—it means tasks within those roles can be
automated. But that's still a massive number representing real people, real
families, real communities.

The Historical Perspective

Here's the stat that should give you hope: About 60% of today's workers are in
occupations that did not exist in 1940.

Think about that. Data scientist. UX designer. Social media manager. Cloud
architect. Cybersecurity analyst. Podcast producer. None of these jobs existed
85 years ago.

By Goldman Sachs' calculations, more than 85% of employment growth since
1940 is explained by technology-driven creation of new positions.

Each wave of innovation—electricity, automobiles, computers, the
internet—eliminated some jobs but spawned others, often in ways that were hard
to predict in advance.

The Messy Transition Period

Here's where it gets uncomfortable. While history suggests long-term job
creation, the transition period is painful:

Goldman Sachs economists forecast:

  • Unemployment might tick up by about 0.5 percentage points during the AI
    transition
  • Roughly 6–7% of the U.S. workforce might be displaced in the short run
  • Within a couple of years, the labor market historically adjusts and overall
    employment recovers

"A couple of years" sounds quick on a historical timeline. It feels eternal when
you're unemployed and can't pay rent.

What's Often Missing from Optimistic Narratives

The "don't worry, new jobs will emerge" crowd isn't wrong. But they're glossing
over real problems:

1. Geographic and Demographic Disparities
Job losses will be concentrated in certain regions (Rust Belt manufacturing
towns) and among certain demographic groups (routine office workers, customer
service reps).

New jobs may emerge in coastal tech hubs or require skills the displaced workers
don't have. A laid-off accountant in Ohio might not benefit from new prompt
engineering jobs in San Francisco.

2. Skills Gap
Even when new jobs emerge, workers displaced from automated roles may lack the
training to fill them.

"Just learn to code" was already bad advice. It's worse when AI can code. The
transition requires massive, well-funded retraining programs—which most
countries don't have at scale.

3. Political Backlash
Job displacement historically triggers:

  • Protectionist trade policies
  • Union resistance to new technology
  • Political instability
  • Restrictions on AI deployment

If enough people get hurt fast enough, democratic societies will slow AI
adoption through regulation. That's not speculation—it's pattern recognition
from previous technology transitions.

4. Quality of New Jobs
Not all new jobs offer comparable wages or working conditions to those they
replace.

The industrial revolution created factory jobs—but early factories were
dangerous, exploitative, and low-paying. It took decades of labor organizing to
make them decent.

Will new AI-era jobs be good jobs? Or will we see a race to the bottom?

Emerging Roles

We can already see outlines of new employment categories:

AI-Adjacent Roles:

  • Prompt engineering and AI interface design
  • AI ethics and safety specialists
  • AI system auditors and validators
  • AI maintenance and fine-tuning technicians
  • Human-AI collaboration specialists across fields

Roles AI Enhances Rather Than Replaces:

  • Teachers (AI tutors assist, teachers mentor)
  • Doctors (AI diagnoses, doctors treat whole patients)
  • Lawyers (AI researches, lawyers strategize)
  • Therapists (AI schedules, therapists connect)
  • Managers (AI analyzes, managers lead)

Entirely New Categories (We Can't Predict Yet): Just as the internet created
"social media manager" and "SEO specialist" and "podcaster," AI will create
roles we haven't imagined. This is the 60% of jobs that didn't exist in 1940
pattern repeating.

The Realistic Timeline (2025-2035)

Here's what the next decade likely looks like:

2025-2027: Early Disruption

  • Customer service, data entry, basic content creation see first wave of
    automation
  • Tech companies implement AI faster than other sectors
  • Unemployment in affected sectors ticks up 1-2%
  • Political pressure builds for AI regulation

2027-2030: Transition Chaos

  • Broader automation reaches mid-level office workers
  • Retraining programs scale up (slowly)
  • Some displaced workers find new roles, many struggle
  • Regional disparities become politically explosive
  • New job categories emerge but don't fully replace losses yet

2030-2035: Stabilization

  • Labor market adjusts, overall employment recovers
  • New roles now account for significant job growth
  • Education systems start producing workers for AI-era jobs
  • Safety nets and retraining programs mature (in countries that invested)
  • Persistent unemployment in regions/demographics that didn't adapt

How to Navigate the Transition

If you're worried about your job, here's practical advice:

1. Develop AI-Adjacent Skills Don't compete with AI—learn to work with it.
Become the person who knows how to get the most out of AI tools in your field.

2. Double Down on Uniquely Human Skills

  • Relationship building
  • Creative strategy
  • Ethical judgment
  • Complex negotiation
  • Emotional intelligence

These are hard to automate and valuable across many roles.

3. Build Transferable Skills The specific job may change, but
problem-solving, communication, and learning ability transfer across roles.

4. Stay Financially Flexible Keep emergency funds. Minimize fixed costs.
Build multiple income streams. The transition period rewards flexibility.

5. Participate in Governance Vote for politicians who take workforce
transition seriously. Support policies like:

  • Robust retraining programs
  • Portable benefits not tied to specific employers
  • Stronger safety nets
  • Progressive taxation to fund transition support

The Key Question

The real question isn't "will jobs exist?" It's "can we manage the transition
humanely?"

That depends on:

  • How fast AI capabilities improve
  • How quickly displaced workers can retrain
  • How generous social safety nets are
  • How much political will exists for transition support
  • How evenly gains and losses are distributed

History suggests we'll have jobs. History also suggests the transition will
hurt—a lot—unless we actively manage it.

The Bottom Line

You'll probably have a job in 2035. But:

  • It might not be the one you trained for
  • The transition might require retraining
  • Some regions/demographics will be hit harder than others
  • Political and social disruption are likely
  • Success depends on adaptation speed and support systems

The messy middle is not an argument against AI. It's an argument for
realistic planning, strong safety nets, and proactive retraining.

Technology doesn't determine outcomes—our collective choices about managing
change do.

Explore the full analysis:
The Second Renaissance: A Balanced Look at AI's Transformation - Complete
picture of AI's transformation timeline

Understand the context:

Build resilience: