Michael Judin | Partner | Judin Combrinck Inc | mail me |
In the seventeenth and eighteenth centuries, one of the most revolutionary ideas in human history emerged – the corporation.
Today, the concept seems entirely ordinary. Companies own assets, enter into contracts, raise capital, employ people and continue operating long after their founders have passed away.
But when the idea first appeared, many regarded it with suspicion and even hostility. How could an artificial entity possess legal rights? How could something that was not a human being own property? How could an organisation exist independently of the individuals who created it?
Religious authorities, political leaders, and established interests questioned the legitimacy of these new institutions. To many, corporations appeared to be dangerous legal fictions that threatened existing social and economic structures.
Yet history delivered a different verdict.
A radically new reality
The great English trading companies demonstrated something profound – human prosperity could be dramatically expanded by creating new forms of organisation. Capital could be pooled. Risks could be shared. Investment could scale. Innovation could accelerate. Shareholders owned. Directors governed. The corporation acted.
The result was an explosion of commerce, investment, innovation and economic growth that transformed the world.
Today, we are witnessing a remarkably similar debate.
The question is no longer whether corporations should exist. The question is whether autonomous AI agents should be allowed to become meaningful economic actors.
Many argue they should not. The concerns are understandable. Every transformational technology creates uncertainty, and every new institution introduces risks.
But history teaches us that society often mistakes the unfamiliar for the unacceptable.
Intelligence – the new capital
The more important question is not whether AI agents will emerge. They already are.
The more important question is whether we will build the legal, regulatory and governance frameworks needed to harness their potential.
Just as corporations enabled capital to operate beyond the limits of individual human lifetimes, AI agents may enable intelligence to operate beyond the limits of individual human attention.
Imagine AI entities capable of running businesses continuously across time zones, negotiating contracts, managing supply chains, conducting research, analysing markets, allocating resources and creating value at a scale previously unimaginable.
What the corporation did for capital, AI agents may do for intelligence.
Governance will create trust
Critics immediately raise an important question – Who governs the AI agents?
History again provides the answer. When corporations emerged, society did not simply grant them power and hope for the best.
We created governance. Boards of directors. Fiduciary duties. Independent audits. Reporting standards. Legal liability. Regulatory oversight. Stakeholder rights. Courts with the power to impose sanctions.
Corporate governance became one of the greatest institutional achievements in economic history, enabling society to capture the benefits of large organisations while mitigating their risks.
The same principle will apply to AI agents. The future will not be a world of ungoverned digital entities operating without accountability. It will be a world of AI governance.
Every significant AI agent will have identifiable sponsors, human fiduciaries, audit trails, compliance obligations, operating mandates and regulatory supervision.
Indeed, AI agents may ultimately prove easier to govern than human organisations. Every decision can be recorded. Every action can be logged. Every transaction can be audited. Every deviation from policy can be identified in real time.
Scaling intelligence
The directors of the eighteenth century governed human organisations. The directors of the twenty-first century may increasingly oversee networks of specialised AI agents operating under clearly defined rules, measurable objectives and continuous oversight.
Humans will continue to provide accountability, ethics, judgement and strategic direction. But increasingly, digital intelligence will execute.
This is not unprecedented. It is the next step in a centuries-long evolution of economic organisation. Argentina offers an important modern lesson, as President Javier Milei recently announced the creation of a new legal category for non-human corporations.
For decades, excessive regulation, economic intervention and institutional rigidity constrained investment, entrepreneurship and growth. Recent reforms reflect a different philosophy – when productive activity is freed, innovation follows, confidence returns and economic dynamism re-emerges.
History repeatedly shows that prosperity grows when institutions evolve to support new forms of value creation rather than resist them.
Resistance rarely retards progress
The corporation was once considered radical. Global capital markets were once considered dangerous. The internet was once considered uncontrollable.
AI agents now occupy the same position. Future generations may look back on today’s debate with the same curiosity we reserve for those who once opposed the corporation itself.
They may ask a simple question – If society ultimately embraced legal entities that amplified capital, why was it so difficult to embrace digital entities that amplified intelligence?
The lesson of history is not that innovation eliminates governance. The lesson is that every major innovation eventually leads to better governance.
We did not stop corporations because they created new risks. We governed them. We are unlikely to stop AI agents for the same reason.
The corporation was the institutional technology that scaled capital. AI agents may become the institutional technology that scales intelligence.
The winners of the next century will not be those who resist this transition. They will be those who build the governance frameworks that make AI trustworthy.
And history suggests that once such a transformation becomes economically valuable, resistance may delay it – but it rarely stops it.
