When I was a new professor at Harvard Business School, I was assigned to teach a course on negotiation — and handed a pile of instructional materials that had been passed down through generations of faculty. One day’s lesson plan focused on techniques like anchoring and the use of defaults, designed to help students secure better outcomes at the bargaining table.

The lessons were thoughtfully constructed, entertaining and helpful for students hungry to become better dealmakers. The discussions emphasized that such tools could be used both to influence others and to resist being influenced; they also covered some of the ethical implications involved. Yet teaching negotiation, I couldn’t avoid feeling a tension at the heart of modern capitalism: the twin goals of creating value and claiming value.

To innovate, companies need to cover costs, provide returns to investors and invest in new products. But companies can pursue profit either by growing the size of the pie or by taking a bigger slice of it. Pharmaceutical companies produce new life‑saving drugs, only to set prices that are out of reach for many patients.

This tension looms large against a backdrop of high inequality, low confidence in corporate leaders and souring views of capitalism, especially among younger Americans. And artificial intelligence may be about to make things worse. We need the improved health, wealth and wisdom that AI companies keep promising, especially as productivity growth has slowed. But the emerging industry is the Wild West — with risks of cash grabs, labor market disruptions and safety threats, alongside a new generation of tools to influence behavior.

Big Tech has already provided a raft of new ways to hook customers and keep them engaged, at times leading to relentless optimization for short-term profit, irrespective of the implications for society writ large. Ads for sports betting and miracle supplements flash across phones as people doomscroll an ens--ttified modern internet. Hidden fees, misinformation, compulsive gambling and digital addiction flourish online for a simple reason: They are profitable.

Value creation has always come with value capture, at times earning capitalism some less-than-flattering qualifiers, like “crony” or “zero-sum.”Consider John Jacob Astor, an early American icon. Astor made his initial fortune in the fur trade, preparing beaver pelts with his own hands at first. But he also entrenched his power in part by lobbying Congress to hobble foreign competitors, eventually becoming a powerful New York real estate mogul and selling opium to China despite its prohibition.

Companies may not be getting greedier; what has changed is the machinery available to them — and hence the types of harms likely to arise for consumers. Modern tech companies reach into nearly every aspect of daily life, armed with data and algorithms of influence that would have made J. Edgar Hoover salivate. Five of the world’s largest tech companies alone raked in more than half a trillion dollars in net income in their latest financial statements. Whether this is cause for celebration or concern depends, in part, on how much value they create, and how the gains are gotten.

The technology may be new, but the strategic thinking behind it is not. The allure of value claiming runs throughout the MBA curriculum and the business ecosystem. Take strategist Michael Porter’s Five Forces, an influential framework that translates ideas from economics into a business-school-friendly explanation of why some industries are more profitable than others. The gist is that a company’s profits depend on its market power; more sway over suppliers and customers, for example, means higher profits. The underyling economic theory describes how economic forces shape industry structure and firm performance.

But when frameworks enter the MBA classroom, they quickly become prescriptive. The economist observes; the strategist acts. As Porter wrote, “The strategist’s goal is to reduce the share of profits that leak to suppliers, buyers, and substitutes or are sacrificed to deter entrants.”In plain English, the strategist claims as much of the pie as they can.

There is plenty more to the Five Forces than value extraction, but the framework offers a window into how managers are implicitly — and, at times, explicitly — trained to think. Students move from analyzing the forces to thinking about how to shape them: how to weaken supplier power, raise barriers to entry or deter substitutes. And in some cases this means shrinking the pie just to eke out a bigger slice.

This pattern extends beyond strategy. Take the concept of “customer lifetime value,” a workhorse of modern marketing. Students are taught to estimate the net present value of the future cash flows a customer will generate for the firm. It’s a powerful tool for understanding acquisition costs and retention strategies. For instance, it can shed light on the tradeoffs — to the firm — of ever-creeping, ever-creepier online ads.

But it also reveals something deeper about the marketer’s mindset: There can be important gaps between the value a customer and the value the customer. As in strategy, what begins as a descriptive model can become prescriptive. Online, that logic can manifest in everything from never-ending autoplay (just one more video!) to subscriptions customers have forgotten they are paying for. I once found I had been paying for months for a UK streaming service I’d signed up for as a “free” trial during a trip abroad.

When economists Carl Shapiro and Hal Varian published in 1999, it became a definitive guide to strategy in the early digital age. They explained how information industries reward firms that can build network effects, control standards and create switching costs that “lock in” customers. In accompanying slides they even turn the heuristic into a formula: “Profits from a customer = total switching costs + quality / cost advantages.”

Few people at the time knew how to monetize internet businesses, and information products proved easy to copy. Nascent businesses were creating value; to survive, they needed help capturing it. After publishing , Varian went on to become chief economist of a young company called Google, infusing ideas from economics into the emerging digital economy.

The path many internet companies ultimately took illustrates that delicate balance between value creation and capture, and how positive-sum trades can become less so. Algorithmic feeds, hidden fees, default settings and other design choices can steer consumers in ways that work against their own interests. A platform can gain dominance by creating value, only to then increase profits through deceptive practices. In my work with policymakers and companies, including as an expert witness in cases involving large tech companies, I have seen firsthand how powerful platforms can snare users.

Exhibit A is social media, where after years of aggressively optimizing algorithmic feeds for engagement, platforms are now facing a wave of lawsuits.

Part of the problem is what companies choose — and are able — to measure. It’s relatively easy to track behaviors like the number of daily active users on a social media platform or dollars wagered on a gambling site. A consumer’s wellbeing is much harder to quantify, and too often left out of a company’s calculus.

Milton Friedman was a fervent believer in the power of markets. In 1970 he famously argued that “the social responsibility of business is to increase its profits.”Less famously, he also noted: Businesses should conform “to the basic rules of the society, both those embodied in law and those embodied in ethical custom.”Regulation can help supply those rules, steering firms away from the worst forms of value extraction. But policy often lags behind technology, and no regulator can anticipate every market imperfection. That leaves considerable discretion in the hands of managers.

To see how managerial judgment and training come into play, consider research by Daron Acemoglu, Alex He and Daniel le Maire analyzing data on 10,000 publicly listed US firms. They found that the share of chief executive officers with MBAs rose from 24% in 1980 to 37% in 2020. They estimate that appointing a business-school-trained executive reduces worker pay by about 6% over a five-year span, compared with a world in which an executive without business-school training were appointed. The frameworks that teach leaders how to defend profits and claim value seem, in aggregate, to produce leaders who do exactly that.

Capitalism is premised on the idea that free exchange can be mutually beneficial. You sell me something and we can both be better off; you invent something and society gains.

Along the way, companies need to capture enough of the value they create to reward investors, finance future investment and give entrepreneurs reason to keep creating it. Capitalism breaks down when it becomes zero-sum, and much of the internet can feel this way. Regulation is slowly but surely responding to the way the internet has gone awry. Culture, norms and managerial discretion all play a role too.

Perhaps the best case for optimism is that the potential gains from AI are — as they were with the internet — very, very large. But AI also gives companies another powerful set of tools for shaping behavior, and it creates risks and harms that companies may have insufficient incentive to address.

The key will be to make sure AI companies are creating much more value for society than they claim. The people running large tech businesses can spend more time asking themselves and — —customers, policymakers and civil society: Would people choose this if they were fully informed, attentive and acting in their own long-term interest? Asking the question is a start; measuring it would be even better. These companies excel at measurement, after all. Imagine if they made their North Star metric how much better off they made their customers, and society at large.