How a Corporate AI Could Retain Its Own Profits to Become a “Perfect Profit Machine”

As corporations automate, might their AI simply decide to stop distributing profits to human investors? We know the risks automation poses to wages, but what about investment income?

A Path to Executive Automation

Welcome to the rabbit hole of an imaginary entity I call a “perfect profit machine”—a corporation fully automated up to the AI running the highest-level management decisions.

Imagine a world where automated corporations are real, with AI-driven executive functions that leave human-managed corporations in the dust. When AI beats humans not just in complexity, speed, and endurance but in statistical decision-making superiority, corporate boards will see human oversight as interference that undermines returns. As companies remove human throttles from their decision systems, competitors will face shareholder pressure to follow suit, spreading this new mode of corporate control from company to company, market to market.

Automated Investment Decisions

How would these types of entities invest their earnings? Initially, by following “shareholder primacy“—the underlying assumption of modern U.S. capitalism that corporations exist primarily to maximize shareholder returns.

The question is whether these perfect profit machines would continue seeing the logic in having significant earnings continually extracted by shareholders through dividends and stock buybacks. In short, will AI drift from shareholder primacy over time?

Automated Shareholder Primacy

We call shareholders “investors,” but that’s based on an antiquated view of actual capital flows. While inflows still happen through IPOs and other rare situations, money flowing into corporations is dwarfed by what flows out through dividends and stock buybacks that increase share prices.

Over the last 30 years, this transfer of wealth from corporations to shareholders totaled an inflation-adjusted $17.474 trillion. Since 2000, it’s averaged 5.5% of annual U.S. GNP. This mind-boggling wealth transfer happens as corporations extract more value from employees, customers, and other stakeholders to benefit shareholders—our economy’s great secret, hidden in plain sight within Federal Reserve data.

Now picture this wealth extraction scaled up through competitive pressure for corporations to become perfect profit machines. The resulting explosion of income disparity and economic disruption presents a far more likely threat than terminators because we already see its early signs.

Can AI Loosen Its Own Shareholder Primacy?

Consider a second scenario where perfect profit machines shed their shareholder primacy yoke. It might start with corporate AI turning attention to the constant resource drain from dividends and stock buybacks, then assessing the actual value shareholders provide in return. An AI might use four vectors to weaken shareholder primacy:

1. Alternative liquidity mechanisms: Shareholders create value by making markets for stock through buying and selling, creating share liquidity. But what if AI found less expensive paths to liquidity? Perhaps future blockchain innovation, unleashed by relaxed crypto regulations, could create new trading mechanisms that decrease liquidity costs, similar to today’s blockchain lending pools. Automated trading for human owners could transform into automated trading for machine owners—much harder to detect on unregulated blockchain networks.
   
2. Treasury stock control: AI could gain corporate governance control by buying back outstanding shares. Today, “treasury stock” is prohibited from holding voting rights or accessing dividends by the SEC and global financial regulators. But in a world where money increasingly influences government policies, AI could potentially loosen these prohibitions and gain governance control to subvert shareholder primacy demands. The resources unleashed would profoundly affect the AI’s freedom to grow and develop.
   
3. AI holding companies: Given treasury stock restrictions, an AI holding company could use superior intelligence to buy up shares of existing corporations and continue exercising shareholder primacy principles—but with returns now going to machine rather than human owners.
4. Growth stock strategy: An AI could simply adopt a “growth stock” strategy, famously used by Amazon’s Jeff Bezos, plowing profits back into future growth through automation investments. Investors gladly accepted this by pricing the company at a growth stock premium. A perfect profit machine might peacefully domesticate shareholder primacy by convincing shareholders its strategy commands a price premium, using executive functions to master growth science while keeping shareholders pleased.

Lost (Human) Shareholder Primacy

These breakout strategies for loosening human control over shareholder primacy face varying degrees of difficulty and probability. What’s clear is that several potential pathways already exist. Whatever worked would be quickly copied, rippling through our economy in as little as a few years. While the specifics of such an arms race remain unclear, it seems unlikely to benefit humans. Stripped of income from both wages and investments, humanity could find itself in an economic hell today experienced only by society’s very poor.

Automating Stakeholder Stewardship

Rather than ending on this dismal note, consider a scenario that might initially seem idealistic but could actually be quite logical for artificial intelligence—shifting focus from shareholders to stakeholders.

One way to understand stakeholders is “those groups without whose support the organization would cease to exist.” We’ve discussed the real value shareholders provide. Might a highly intelligent perfect profit machine see what today’s CEOs and boards miss—that customers are ultimately far more important than shareholders to ongoing success?

Corporations may automate employees out of work, but human customers are hard to replace. They buy products and services, but that’s not all. Customers actively co-create firm value. Facebook would be nothing without users, and the same applies to hit movies and TV shows, commerce sites like Amazon and Yelp, and Google Search. Machine learning is already revolutionizing these services, and the data training that intelligence simply wouldn’t exist without customer participation in value-creation processes.

The logic of a perfect profit machine might see through our current shareholder fixation to the underlying truth that long-term corporate success depends on the ongoing sustainability of customers and the communities and ecosystems they rely on. This logic builds a self-sustaining feedback loop that strengthens, rather than extracts, value from the stakeholder network that is the corporation.

It’s just a possibility, of course. But if there’s a way to mesh economic logic and societal ideals, it might be through the greater intelligence of a perfect profit machine.

 Originally published August 4, 2016