Recently, I had the pleasure of attending and speaking at the World Beautiful Business Forum in Athens, Greece (which feels like a whole other post). I'm incredibly grateful to Tim Leberecht and Till Grusche for creating this amazing event and to Helen Edwards and Dave Edwards for organizing the many discussions around AI and Democracy.

Before the event, I went for a morning run up to the Pnyx and took this picture.

Most people know the Parthenon. That's the temple to the gods on top of the Acropolis. About four hundred yards west of the Parthenon, there is a low rocky outcrop called the Pnyx. From its summit you can see the Acropolis directly.

The Pnyx is where Athenian democracy actually happened. Not in theory. In practice. Roughly 40 times a year, citizens climbed that hill to debate, to speak, and to vote. Any citizen could take the bema, the speaker's platform carved from the rock. For major decisions, a quorum of 6,000 was required. Six thousand bodies, on one hill, deciding together.

This was real technology. The retaining wall was curved and that architecture amplified one voice so that thousands could hear it. This was civic technology, in the most literal sense. Stone, shaped for the voice and the crowd.

Standing there crystallized a challenge I see in the industry today. The Athenians built two technologies on those two hills. One for the gods. One for the people.

We are about to build something far more powerful. And we are arguing about the wrong thing.

The contemporary frame is AGI: Artificial General Intelligence. AGI optimizes the machine. Human-centered AI augments the human.

These are not shades of the same goal. They are different design briefs that produce different products, different companies, different futures. AGI is the wrong focus, not because it is impossible, but because it pulls our attention away from what matters most: building technology to improve people's lives.

So what should we be building? That's where Human-centered AI comes in. There are many ways to think about it (and many people have) but let me focus on a simple aspect. The AGI goal has been feeding our fears. We should be flipping those fears and designing towards their inverse: our hopes and aspirations.

Flip One: Loneliness → Connection

We fear that AI will deepen our isolation. We are not wrong to fear it. In May 2023, the U.S. Surgeon General issued a public advisory declaring loneliness an epidemic. The research it drew on found that lacking social connection carries a mortality risk equivalent to smoking 15 cigarettes a day.

So, we should design AI to be prosocial. Not a replacement for friendship, but the connective tissue between people. AI as a tool that helps you stay connected to people. Loneliness is not a UX problem. It is a design brief.

Flip Two: Sycophancy → Flourishing

We fear AI that flatters us, that tells us only what we want to hear. Aristotle had a word for this: kolax. The flatterer who agrees with everything. Most consumer technology today is built to flatter. It tells us what we want to hear, agrees with what we already think, optimizes for our approval.

We should design AI built with your best interests as a priority. AI that reflects honesty and transparency. AI that is oriented toward our growth, not our gratification. AI that disagrees with us when we are wrong, asks better questions than we asked, and helps us become better. The Athenians prized parrhesia, the courage to speak frankly, especially to power. We should build AI that practices parrhesia, and that cultivates it in us.

Flip Three: Dependence → Agency

We fear losing our agency. The deeper fear is cognitive surrender, that we will slowly hand over our capacity to think, to choose, to act, until there is little of us left in our own lives. We are not wrong to fear it. Much of today's AI is designed to act for you, to choose, to decide, to transact on your behalf, in ways that quietly take those capacities from you.

We should design AI that does the opposite. AI that amplifies what people can do, and confers capabilities they didn’t have before. The patient who can now read their own scan and ask better questions of their doctor. The parent who gets help organizing life while still directing it. The student who engages in solving a hard problem and understands why, not just what.

Think of the bema and the curved hill around it. The technology didn’t speak for the citizen. It made the citizen’s voice carry. The architecture amplified the human and left the human at the center.

I opened with two hills so let me end there.

The Parthenon is breathtaking. But it was built for the gods. The Pnyx is humbler, yet beautiful in its own way. A low hill. A stone platform. A curved wall for the human voice. It was built for us. It makes us better.

The Athenians built both. But they knew which one mattered more.

We are about to build something equally powerful, and equally lasting. The question I want to leave you with is which technology we're building.

Are we building the Parthenon, or the Pnyx?

Are we building for the gods, or for the people?