The argument that changes how your audience thinks about AI — permanently.

We talk about AI as a technology story. Michael argues it is something older and more fundamental: a property rights story.

Every transformative economic leap in history — from the Magna Carta, to the joint-stock company in 17th-century Amsterdam, to Deng Xiaoping's transfer of home ownership to 800 million Chinese — was triggered by the same event: a property right was extended to a vastly larger class of people. Each time, the economic growth and human flourishing that followed was beyond anything the prior era could have imagined.

We are at that moment again. Data — specifically, the data generated by human beings going about their lives — is the most valuable commodity of this era. And right now, almost none of it is owned by the people who created it.

The AI age is producing two crises that look separate but are the same problem wearing different clothes.

The first is a crisis of personal meaning: as AI automates more of what humans do, the question of what it means to contribute — to work, to create, to matter — becomes urgent and personal. Audiences feel it in their bones but rarely hear it answered convincingly from a stage.

The second is a crisis of economic value creation: for companies of every size, the ability to generate sustainable revenue in an economy increasingly shaped by a handful of data-aggregating platforms is genuinely under threat. The Magnificent Seven of Silicon Valley are not simply big companies — they are an architectural layer capturing the returns that should flow to the businesses and individuals who actually create the value.

Michael's argument is that these two crises have the same solution.

The answer is not regulation. It is not slowing down. It is architecture.

Using privacy-preserving technologies — zero-knowledge proofs, confidential computing, verifiable digital identity systems, decentralized data infrastructure — it is now technically possible to give individuals and the businesses they represent genuine, verifiable control over their own data. Not as a policy aspiration. As an engineering reality.

This is what Michael means by Human Agency in the AI Age: not a slogan, but a specific technical and governance design choice being made right now, by engineers and companies and policymakers who mostly don't know they are making the most consequential architectural decision in the history of the internet.

Michael has been in the room for many of those decisions. He makes the case for the right ones — with the clarity of a journalist, the rigor of an economist, and the conviction of someone who has seen the alternative up close.

The emergence of AI agents makes this argument urgent in a specific, time-sensitive way.

An AI agent is supposed to act on your behalf — to know your preferences, protect your interests, and work for you. That is the promise. But there is a question buried inside that promise that almost no one on a mainstream stage is asking: whose agent is it, really?

If the AI agent that you, or your business, use is controlled by the platform that built it — optimized for the platform's engagement metrics, the platform's data harvest, the platform's revenue — then it is not your agent in any meaningful sense. It is a tool of surveillance dressed up as a personal assistant. This is what Michael calls digital feudalism: a system where the lords are the platforms and the rest of us are serfs who happen to use nice interfaces.

The alternative — AI agents cryptographically constrained to serve the person or business that employs them, whose data is verifiably private, whose decisions are genuinely fiduciary — is not science fiction. It is being built. And the companies that understand this now will be positioned to capture the extraordinary economic upside that follows from genuine human trust.

Michael is not a critic of technology. He is not a critic of business. He is a critic of bad architecture — and an evangelist for the economic windfall that better architecture unlocks.

If we extend genuine data ownership to individuals and the businesses they work with and for, we do not shrink the economy. We expand it in ways that are historically without parallel. The locked value in consented, privacy-preserving, verifiably-human data — data that AI companies cannot currently access because they cannot trust it — is an enormous opportunity for every company in every industry.

This is the message that separates Michael from every other speaker on technology and AI: he brings the optimism, and he can prove it.