In a twist nobody saw coming, liberal arts graduates might be the last ones standing when AI takes over—and the gig workers training their robot replacements could be teaching us the most important career lesson of all.

Teaching AI To Do Your Job Is Also A Job

This Wall Street Journal article shows, gig workers are literally training AI to replace them. A video editor named Katie Williams spent six months teaching an algorithm how to do her job—clips, captions, the works. She taught AI to be as good as her in exchange for a $150 an hour fee that Mercor paid her. That $150/hour rate? It’s severance pay disguised as training wages.

On the Mercer website you find an Interventional Radiologist, Biotech Researcher Scientist, a Medical Researcher and a Banker & MBA student explaining why they are teaching AI to do their job.

The Three Skills AI Can’t Steal (Yet)

A friend and former colleague from my days in advertising recommended Fareed Zakaria’s book “In Defense of a Liberal Education” (2015). Once I started reading it, I could not stop. Here are some ideas that stayed.

Liberal education stresses the importance of writing, thinking and learning. The combination of these three skills makes a person a powerful candidate when AI is taking over jobs.

“Tech training teaches you to fish in a specific pond. Liberal Education teaches you to adapt to any body water.”

Writing teaches thinking. When Zakaria arrived at Yale from Mumbai, he could ace tests but couldn’t express an original idea. His English professor covered his essays in red ink. That pain taught him something no algorithm can replicate: “I realized that in coming from India, I was pretty good at taking tests and regurgitating things I had memorized; I was not so good at expressing my own ideas.” He has certainly more than made up for that.

This observation hits particularly hard in both US and Indian contexts. In India’s education system of the 1970s—and in many institutions today—students chose “streams” at age 16: science for smart kids, commerce for rich kids, humanities for everyone else. The US system, for all its flaws, allowed Zakaria to explore freely. That freedom, not the coursework itself, made the difference.

Speaking teaches persuasion. Facebook’s billion-dollar insight wasn’t technological—it was psychological: people would voluntarily expose their real identities online. That may have something to do with the combination of Ancient Greek in high school and psychology that he studied.

Learning to learn beats learning facts. MIT economist David Autor divides jobs into three categories: bottom-tier service work requiring human touch (eg care giving for seniors), middle-tier white-collar routine (being automated away right now), and top-tier creative problem-solving requiring “creativity, judgment, and common sense—skills that we understand only tacitly. Anyone who has aspirations to take up top tier roles will do well to build an appreciation for Liberal Education and moving away from a vending machine model of Higher Ed where students push in coins to find jobs that return the money quickly.

The Global Talent Arbitrage Collapses

When AI can do routine analysis, the cost advantage of outsourcing evaporates. A company in Boston doesn’t need to hire cheaper analysts in Bangalore when an algorithm costs pennies per query.

The BPO sector that employed millions in India faces an existential threat. Indian professionals who built careers on cost advantage must now compete on creativity, synthesis, and innovation—the very skills liberal education develops.

The same pattern hits the US differently. American workers who assumed their geographic proximity protected them discover AI doesn’t care about time zones. The only moat left is uniquely human value-add.

What Organizations Can Do

Hire Curious Learners

Google’s data showed GPA and test scores barely predicted job performance. “Learning animal” mindset did. In interviews, stop asking about skills. Start asking: “Tell me about something complex you taught yourself in the last year unrelated to your job.”

For companies accustomed to filtering resumes by GPA and college tier, this requires cultural rewiring.

Build T-Shaped Professionals

The vertical bar is deep expertise in one domain. The horizontal bar is 20% depth in 3-5 adjacent fields. Create “discipline crossing” metrics. Track how many people are learning outside their function. Reward it like you reward promotions. A software engineer who understands psychology and design thinking is worth five who only code.

Avoid asking for slide decks for strategic decisions. Instead ask for a 3 page note that explains the key idea, the timelines, the potential flaws, payback time etc.

Create the Reinvention Fund

Give every employee some money every year to take any course in anything. No approval needed, especially if they take “opposite subjects”. For example when the accountant takes creative writing or when the engineer studies philosophy? They’re building the horizontal bar that makes them irreplaceable when AI comes for the vertical one.

Employee retention improves when employers invest in career development. People stay where they grow. Companies that invest in learning see lower turnover even when competitors offer higher salaries.

What Individuals Can Do (Especially in Tech)

The Adjacent Possible Reading List

Rule: 50% in your field, 30% in adjacent fields, 20% completely random.

If you’re in tech, read history, philosophy, and design. If you’re in business, read psychology and sociology. When AI can code better than you, your value is connecting dots between domains AI hasn’t linked yet.

The Daily Writing Practice

30 minutes every morning: Write about anything. Don’t publish it. Don’t polish it. Just think on paper. You can’t write clearly without thinking clearly. And you can’t think clearly without writing.

After three years of this practice, you’ll have clearer thinking than 95% of your peers. That’s called skill compounding.

Be a polymath

Stop thinking “I’m a software engineer.” Start thinking “I’m someone who combines coding + psychology + climate activism.”

That’s a career AI can’t touch.

Build competence through
Deliberate Practice ie real projects, proven work

Build confidence through
Small wins, reframing failure as data, acting before you feel ready

Building AI Proof Skills

Here’s how to know if you’re building AI-proof skills:

• Can you explain your work to a smart 12-year-old? If not, you don’t understand it well enough.

• Can you learn one AI proof skill – the ability to work across cultural contexts.

• Can you learn something unrelated to your job and connect it back? If not, you’re too specialized for an unpredictable future.

Those gig workers training AI aren’t doomed. The ones treating it as a learning opportunity—understanding the algorithm, spotting its weaknesses, asking what it can’t do—they’re building the meta-skill that matters.

The ones just collecting the $150/hour? They’re in trouble. So it depends how they do the same task.

In Summary

The future doesn’t belong to the narrowly trained. It belongs to the broadly educated who can focus. Liberal education is a terrific way to think about the future of work. It lies at the intersection points of multiple disciplines. <here are jobs that need multi disciplinary thinkers>

Not OR. It is an AND world

That’s the only sustainable advantage left.

If you had to choose between your kid becoming the best video editor in the world or someone who understands video editing, psychology, and storytelling well enough to build something nobody’s thought of yet—which would you pick, and why?