We have spent much of the AI governance debate asking what artificial intelligence should be allowed to do. I believe there is a prior question we have not examined closely enough: what makes algorithmic power legitimate? Montesquieu’s warning about concentrated power was written for political institutions, not algorithms. Yet the principle behind it has surprising relevance to AI. When the same organization can effectively design an algorithmic system that fuses rule-making,

or the past seventy-five years, AI warnings have repeatedly gone mainstream while the approach capable of answering them has remained largely invisible. Beneath all these warnings lies one simple issue, always the same: not intelligence, but power. Algorithmic or otherwise, power must be separated, and ultimately encounter power, if it is not to become absolute and irreversible.

Tell me what you buy, and I’ll tell you who you are. Agentic commerce could proceed from that premise. AI agents are increasingly acting on behalf of consumers, with agentic commerce projected to reach trillions by 2030. This shift raises concerns, particularly in one sector that may not be the first to come to mind, beauty and fashion. Optimized to learn from what we desire, what we can afford, what we fear missing, AI agents in commerce are machines built to understand who

More and more voices around the world are acknowledging our algorithmically mediated reality. Reality is to say that algorithmic systems are algorithmic power. The legitimacy we never questioned In March 2023, the Future of Life Institute and thousands of signatories, including Yoshua Bengio, Stuart Russell, Yuval Noah Harari, and Steve Wozniak, called for a six-month pause in training AI systems more powerful than GPT-4. This asks for the power to pause. Later that year, Dar

Gates is wrong to suggest that the world needs institutions capable of governing AI. Any institution created to govern power acquires power of its own and, like any power, tends toward self-preservation and expansion, ultimately organizing itself around its own continuity rather than merely constraining the power it was created to govern. What the world needs first is Trias Algorithmica. Just as Trias Politica separates rule-making, executive, and adjudicative powers, we now

Many organizations implementing AI agents tend to focus too narrowly on a single decision-making model, falling into the trap of assuming a one-size-fits-all decision-making framework, one that follows a typical sequence in any circumstance: from input to research and analysis toward decision, then execution, eventual evaluation, and hopefully, lessons learned. However, it oversimplifies reality.