Giorgos Tsetis, the founder of hair growth supplement company Nutrafol who now invests in artificial intelligence ventures, is calling for executives to be held personally liable when safety failures cause harm — a standard, he argues, long applied to consumer products but absent from the AI industry.

In a commentary published by Fortune on October 17, 2026, Tsetis recalls leading Nutrafol for almost a decade, during which he was responsible — not just morally but legally — if the supplements his company sold caused harm. "It kept me awake at night. And that was a good thing," he writes, describing personal legal responsibility as "a powerful tool" that pushes corporate leaders to take product safety seriously and gives consumers confidence. By contrast, he says, such accountability is "not currently something causing AI leaders to lose sleep. It should be."

Tsetis acknowledges that some believe an added layer of accountability would stifle innovation, but says his experience building Nutrafol into what he describes as a multi-billion-dollar company points the other way. In the supplements industry, the law has long held those in charge personally responsible for the safety of what their companies sell — an accountability he credits as essential to building the consumer trust his business needed to succeed.

Today, Tsetis invests in AI and AI-enabled companies through Great Things™, a new investment model, and through his family office. His investments include frontier AI companies such as Anthropic and xAI, infrastructure and compute-related companies such as Positron AI, and AI-enabled science and health platforms such as Lila, which uses AI to accelerate discovery. He says he is putting capital behind these companies because he wants them to succeed and believes AI, used safely, can help solve some of humanity's toughest problems.

But as an investor within the ecosystem that makes AI innovation possible, Tsetis says he has observed a gap: personal legal accountability. Executives are currently shielded behind their companies, with corporate financial penalties for misconduct or negligence often amounting to a fraction of turnover and valuations — "just another line item." "Responsibility that stops at the corporate entity isn't responsibility. It's an expense," he writes. Meanwhile, AI leaders can make decisions from which they personally stand to make billions while risking nothing themselves if their mistakes expose the public to danger — an asymmetry he calls indefensible.

Tsetis notes that AI rules are evolving quickly. States across the US are introducing their own AI laws, and Congress is weighing legislation that would impose safety requirements on frontier AI companies. With those federal proposals still taking shape, the emerging regulatory landscape is one of state measures alongside pending national rules — a mix that leaves open the question at the center of his commentary: who, specifically, answers when AI systems fail. At the end of September, leaders of several of the biggest AI companies went to the White House and signed an accord committing to internal risk controls, independent audits and board-level oversight of safety. He describes these steps, along with the safety frameworks leading labs have published, as welcome — but stresses that the frameworks and the accord are voluntary, and that "when it comes to AI, voluntary isn't enough." "Nothing yet holds the people at the top personally legally accountable," he adds.

The result, he argues, is a legal system in which companies mostly pay but the people who lead them rarely do. An AI executive who misstates revenue to investors can be held personally responsible, yet decisions that could contribute to anything from an individual suicide to large-scale cyberattacks and geopolitical destabilization are made without personal legal jeopardy. Tsetis says the tech industry has already seen where that leads in social media, and that with AI the stakes are potentially much higher.

Drawing on his experience building Nutrafol, Tsetis writes that personal liability changes how a business is run. A safety checklist tells a company what to test; personal exposure makes its leader worry about what is not on the list. Under that pressure, executives drive their teams to identify risks, test for them and put safeguards in place before problems emerge, ensure that problems reach them early rather than being smoothed over on the way up, and accept slowing down or saying no — even when money or a competitive edge is on the table — because their name is attached to what happens next.

Tsetis concedes that asking for personal liability for AI leaders is a significant step, but argues AI is unlike any other industry. A supplement reaches only the people who choose to buy it, while AI is becoming embedded in everyone's lives — whether boarding a plane or turning on the lights. "We are all 'customers' of AI's products," he writes. AI technologies are built into the systems society depends on, from power grids to hospitals to the classrooms where children learn, and a single failure could ripple across a city, a country or beyond. Unlike almost any other product, he adds, the people building AI have warned it may one day escape human control. "Extraordinary risks call for extraordinary accountability."

To be clear, Tsetis says, he is not arguing that AI executives should be punished for every failure. Just as he could not guarantee that every one of his products was perfect, no one can guarantee an AI system is safe. But he argues there should be consequences — not just for the company but for the individuals behind the decisions — when leaders know their systems pose serious risks and ship them anyway, when they inadequately fund their safety teams or ignore their warnings, or when they hide problems from regulators and the public.

Whether that accountability is civil or criminal is not for him to say, he writes, nor are questions of where thresholds sit, which harms count and who audits what — matters he leaves to legislators, lawyers, technologists and ethicists. But whatever obligations they land on, he argues, a named senior executive should be personally responsible for meeting them. "When the people at the top know the risks, what they do next should carry real personal stakes." Whether US lawmakers ultimately write personal accountability into binding law, and in what form, will be among the clearest markers to watch as the state and federal measures now in motion take shape.

Tsetis concludes that he believes AI has the potential to transform everyone's lives for the better while sharing the common concern about its dangers. Personal legal responsibility, he says, focuses minds on safety, builds public trust and helps ensure that the race among AI leaders to come out on top does not come at everyone else's expense. It is a standard, he argues, long applied to those making products that go into human — and it must now be applied to those making products that could shape the future of humanity itself.

As Fortune notes, opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune. The original commentary is available on Fortune.com.