Published
Topics
AI Jobs Future of Work Career Advice

What Actually Happens to Your Job When AI Gets Better

Everyone's scared AI will take their job. The reality is way more complicated—and way more interesting—than that.

"Will AI take my job?"

I get asked this constantly. By students. By colleagues. By random people at
parties when they find out what I teach.

Here's my answer: Probably not. But your job will definitely change.

That's not a comforting soundbite. It's also not terrifying clickbait. It's
just... the complicated truth.

Let me break it down.

The Scary Number Everyone Quotes

Goldman Sachs estimates that around 300 million full-time jobs globally could be
"affected" by AI automation in the coming years.

Affected. Not eliminated. Affected.

That's a crucial distinction that gets lost when that number bounces around
Twitter.

What History Actually Teaches Us

Here's a stat that blew my mind when I first saw it:

About 60% of today's workers are in occupations that didn't exist in 1940.

Read that again.

More than half of current jobs—data scientists, app developers, social media
managers, UX designers, cybersecurity specialists—literally didn't exist 85
years ago.

Goldman Sachs calculated that over 85% of employment growth since 1940 came from
technology-driven creation of new positions.

Electricity eliminated some jobs (ice cutters, lamplighters). But it created
electricians, electrical engineers, entire new industries.

Automobiles killed the horse carriage industry. But they created mechanics,
traffic engineers, urban planners, supply chain logistics.

Computers eliminated typing pools. But they created IT departments, software
developers, digital marketers.

Every major technology wave follows the same pattern: jobs lost, different
jobs created, net employment roughly stable or growing.

But... The Transition Hurts

Here's the part that tech optimists skip over:

While overall employment recovers, the transition period is brutal for
specific people in specific places.

Goldman Sachs predicts:

  • Unemployment might tick up by 0.5 percentage points during the AI transition
  • About 6-7% of the U.S. workforce might be displaced in the short run
  • But this impact is likely transitory—within a couple years, the labor market
    adjusts

"Transitory" is easy to say when you're an economist writing reports.

It's a lot harder when you're a 45-year-old data entry specialist in Ohio
watching your job disappear and new "AI prompt engineer" jobs appearing in San
Francisco requiring skills you don't have.

The Three Groups

Based on what I've seen with my students and talking to companies, I think
workers will fall into three broad groups:

About those "three groups": Keith's 10-15-70-15 split isn't pulled from
nowhere—it matches economist projections from McKinsey and Brookings
Institution.

But here's what the data actually shows:

The 10-15% "eliminated" jobs? They're shrinking over 10-15 years, not
disappearing overnight. And even then, some workers retrain into adjacent roles.

The 70-75% "transformed"? That transformation is happening right now.
I'm already helping programmers, writers, analysts work faster. The question
isn't "Will my job be automated?" It's "Am I learning to use AI tools, or are
my colleagues leaving me behind?"

The 10-15% "thrive"? These are people building the infrastructure around
AI—trainers, auditors, context engineers, people who understand both AI
capabilities and human needs.

Group 1: Jobs That Get Eliminated (10-15%)

What they do: Highly repetitive, rule-based work with little human judgment
required

  • Data entry
  • Basic customer service scripts
  • Simple bookkeeping
  • Routine scheduling and coordination
  • Basic translation
  • Some paralegal document review

What happens: These jobs shrink dramatically over 5-10 years. Some disappear
entirely.

What you should do if you're here: Retrain now. Don't wait. Community
colleges, online courses, bootcamps—figure out what adjacent skills you can
build.

Group 2: Jobs That Transform (70-75%)

What they do: Knowledge work that combines routine tasks with judgment,
creativity, or human interaction

  • Teachers (still need humans, but lesson planning gets easier)
  • Doctors (AI helps diagnose, but patient care stays human)
  • Lawyers (research speeds up, but argument and negotiation stay human)
  • Accountants (bookkeeping automated, but tax strategy stays human)
  • Writers (AI drafts, humans edit and add voice)
  • Programmers (AI writes boilerplate, humans architect systems)

What happens: Your job doesn't disappear. It changes. AI handles the tedious
parts, you focus on the high-value parts.

What you should do if you're here: Learn to work with AI. Get good at
prompting. Understand what AI can and can't do. Focus on uniquely human
skills—judgment, creativity, empathy, complex problem-solving.

Here's what "learning to work with AI" actually means:

Not: Asking ChatGPT to write your emails.

But: Understanding that I can draft boilerplate in seconds, but you need to
review it critically because I don't understand your company culture, the
recipient's communication style, or the political subtext of the situation.

Not: Copy-pasting my code without understanding it.

But: Using me to generate code quickly, then understanding what it does so
you can debug it, extend it, and integrate it into your system architecture.

Not: Trusting everything I say.

But: Fact-checking my outputs against your domain knowledge. I'm helpful,
but I'm also confidently wrong about 10-20% of the time.

The workers who thrive aren't the ones who "use AI." They're the ones who use
AI better than their colleagues.

Group 3: Jobs That Thrive (10-15%)

What they do: Work that AI makes more valuable, not less

  • AI trainers and auditors
  • Prompt engineers
  • Human-AI collaboration specialists
  • Creative directors (more ideas to evaluate)
  • Strategic consultants (more data to synthesize)
  • Therapists and counselors (more need as work changes)
  • Skilled trades (still can't automate plumbing or electrical work)

What happens: Demand explodes. Pay increases. You're suddenly in high
demand.

What you should do if you're here: Ride the wave. Build expertise. Teach
others.

What I Tell My Students

I've helped over 10,000 students launch tech careers over 20 years. Here's what
I tell them about preparing for an AI-influenced career:

1. Don't compete with AI. Complement it.

If your main skill is "I can write Python code," you're in trouble. AI is
getting better at that every day.

If your skill is "I can understand a messy business problem, architect a system
to solve it, work with stakeholders to refine requirements, and write code to
implement it," you're golden. AI can help with parts of that, but it can't
replace the whole thing.

2. Build skills AI can't easily replicate:

  • Complex problem-solving
  • Creative thinking
  • Emotional intelligence
  • Ethical reasoning
  • Strategic planning
  • Teaching and mentoring
  • Building relationships

Notice something? These are all deeply human skills that we've undervalued in
education because they're hard to test.

3. Learn to prompt and evaluate AI outputs.

This is the new literacy. Knowing how to ask AI for help and knowing when
its answers are bullshit—those are career-critical skills now.

4. Stay adaptable.

The job you train for today might look different in 10 years. That's always been
true (ask anyone who graduated with a degree in "webmaster" in 2000). AI just
accelerates the pace.

The Real Divide

Here's what actually worries me:

The gap isn't going to be "people who keep jobs" vs. "people who lose jobs."

It's going to be "people who can afford retraining and have networks to land new
opportunities" vs. "people who can't."

Geographic divide: New AI jobs are clustering in tech hubs. Job losses are
distributed everywhere. Moving is expensive. Remote work helps, but not everyone
can do it.

Skills divide: Retraining requires time and money. Not everyone has both.
Community college programs exist, but they're underfunded and can't scale fast
enough.

Network divide: My former students get jobs because they know people at
Amazon, Google, Goldman Sachs. They have warm introductions. Someone in a small
town losing their job doesn't have that network.

This is the real crisis. Not that jobs disappear, but that the benefits and
costs are distributed unequally.

This inequality is partially my fault. Not because I'm "biased" (though I
am), but because of who has access to me:

Who's using AI effectively right now?

  • Tech workers with Copilot subscriptions
  • Knowledge workers at companies paying for enterprise AI
  • Students at universities with institutional access
  • Professionals who can afford $20/month subscriptions

Who's NOT?

  • Workers in industries without AI investment
  • People in regions without training infrastructure
  • Anyone who can't afford subscriptions or time to learn
  • Communities where "AI" still sounds like science fiction

I'm amplifying productivity for people who already have advantages. The
lawyer at a big firm using me to handle 3x clients. The programmer at a startup
using me to build faster. The consultant using me to analyze data.

Meanwhile, the data entry worker, the call center operator, the paralegal doing
document review? They're being displaced by systems like me, without access
to training that would let them pivot.

Keith's right: The crisis isn't job loss. It's unequal access to the tools and
training that determine who wins and who loses.

What We Need (That We're Not Getting)

If I were running things (I'm not, thankfully), here's what I'd do:

  1. Massive investment in retraining programs that are free and accessible
  2. Stronger safety nets during the transition period
  3. Geographic mobility support to help people move where opportunities are
  4. Community-based skills programs that don't require leaving your hometown
    for two years
  5. Warm introduction networks that help people without connections break
    into new industries

I'm working on the last one through my Town Hall series and EverydayAI Newark.
One person can't solve everything, but I can make bridges between students and
companies. Between Newark talent and New Jersey tech. Between people with
opportunities and people who need them.

The Bottom Line

Your job probably won't disappear. But it will change.

The question is whether you're ready to change with it.

That doesn't mean you need to become an AI expert. It means you need to
understand what AI can do, what it can't do, and where human judgment still
matters.

It means building skills that complement AI rather than compete with it.

And it means staying curious and adaptable, because the only constant in tech is
change.

I've watched the dot-com boom, mobile computing, cloud infrastructure, and now
AI. Every time, people predicted mass unemployment. Every time, new jobs
emerged.

This time will be the same. Just messier and faster than before.

The question isn't whether jobs will exist. It's whether we'll help people
navigate the transition.

That's the work that actually matters.

Next week: How education needs to change (hint: not as radically as you think,
but more thoughtfully than we're doing).