Will AI replace all of us? Philosopher Bas Haring on the role of AI agents

The breakthrough that wasn’t really a breakthrough

Bas was already studying artificial intelligence in 1988 – at a time when the field didn’t even have a proper name yet. Looking back on the past decades, he says:

“Honestly, not that much has changed since my student days, even though the world looks completely different. What has changed most is that computers have become much faster, which creates far more possibilities.”

For example, his smartwatch is about a hundred thousand times faster than the computer he used during his studies. Technologies like facial recognition were theoretically possible back then too – but processing a single scan would have taken two days.

He’s just as clear about the so-called AI breakthrough: “ChatGPT has put AI firmly on the agenda, but in terms of actual innovation, it’s not as revolutionary as it seems. There hasn’t really been a fundamental breakthrough.”

The takeaway? AI isn’t something mystical or untouchable. It’s mathematics, running on very fast computers.

Will AI make us unemployed?

Short answer: probably not. Bas uses the example of excavators to explain what historically happens when new technology is introduced:

“The invention of the excavator made digging so fast that other parts of the construction process couldn’t keep up. This created more demand for carpenters and bricklayers. Many of the former diggers eventually found work in those roles.”

This pattern keeps repeating. Bas gives a modern example: “In 1988, no one could have predicted that you could make money in the future by filming yourself playing video games. Yet today, that’s a real job.”

So he’s not worried about structural unemployment caused by AI: “When one field changes, something new emerges somewhere else.”

What you automate becomes worthless

This is perhaps the most underestimated pitfall of agentic AI—and Bas explains it with striking clarity using a birthday card:

“A handwritten birthday card is valued by most people. But if you let an AI agent choose, write, and send that card, it suddenly becomes worthless. The value lies in the fact that someone took the time to send it to you.”

Automating cost-heavy processes in your business? Fine. But automate something people perceive as valuable, and you risk destroying exactly what made it valuable.

Bas experienced this himself when writing a foreword for a book on AI: “It’s very possible that readers think the foreword was written by AI. But the meaning of a foreword isn’t the text itself. The meaning is that I’m showing I think it’s a cool book or a cool author.”

His solution? Write a sentence AI could never have produced. And his advice to anyone working with AI agents: “Remember this: when actions are automated away, those actions themselves risk becoming worthless. Keep that in mind.”

The webcam wasn’t intentionally invented either

One of the most fascinating insights about innovation doesn’t come from a tech giant, but from a computer scientist in Oxford. In the 1980s, he pointed a digital camera at a coffee machine so colleagues could see whether there was any coffee left.

A small problem. A simple, slightly quirky solution. A few weeks later, people from Johannesburg, Toronto, and Tokyo started tuning in. That’s how the webcam was born.

Bas’ point: “The webcam wasn’t consciously invented. It emerged because someone was curious and tried to solve a small problem.”

His advice in times of big change: “Instead of trying to do something big and impressive with AI, you can also start small in your own environment. Be curious. Tinker with AI agents. Who knows – you might end up creating something amazing, just like the accidental invention of the webcam.”

So what’s left for us?

With the rise of AI and AI agents, the value of what we do is shifting. Bas discussed this with several general practitioners, including his own.

The analytical part of their work—diagnosing based on information – can increasingly be supported by AI. But the truly valuable part cannot.

As his doctor put it: “I block fifteen minutes in my schedule for someone. That’s what I do. That’s the core of my work, and AI will never replace that.”

Bas strongly believes in a shift, not a disappearance: “It wouldn’t surprise me if, over time, work shifts from being primarily analytical to becoming more relational or social. I think we’re living in a fascinating time where many things will change - but there will always be something for everyone to do.”

* This blog is based on the keynote by Bas Haring, philosopher and professor at Leiden University, during IncentroCon Agentic ’26 on March 18, 2026.