Long-Term Benefit Trust
The Long-Term Benefit Trust is a trustee body at the AI company Anthropic that is gradually allowed to fill a portion of the board seats. It is meant to ensure that long-term safety and the common good have a say in decisions, rather than just investor profit.
A company is normally controlled by a board of directors. That’s a small group of people who can appoint the CEO, oversee them, and if necessary, remove them. Whoever gets to determine this group holds the real power in the company. Typically, it’s the financial backers who fill these seats, meaning the people who bought shares. The Long-Term Benefit Trust is a different approach: an independent group of trustees is allowed to appoint a portion of these board seats. This construction was introduced in 2023 at the American AI company Anthropic, which develops the chatbot Claude.
A safety promise meant to withstand investors
Almost all AI companies say that safety matters more to them than rapid growth. The problem: such promises are non-binding. Once investors have paid billions, they want to see returns. If a competitor releases a stronger product, the pressure grows to release one’s own model faster and with less caution. A mere promise rarely holds up against that pressure.
The Trust is meant to build this promise into the ownership structure itself. The trustees have no share in the profits. So they lose no money if the company grows more slowly. Their explicit mandate is the long-term welfare of the public, not the share price.
How much such constructions really hold up is disputed. In 2023, the then-CEO of OpenAI was dismissed by a similarly nonprofit-minded body and reinstated five days later. Critics have since argued that such a body can hardly assert itself against investors and its own workforce when it really matters. Supporters counter that a formal hurdle is still better than none at all.
How the trustees get their seats
Legally, the Trust is what’s known as a purpose trust. This is a trust fund that belongs to no person but serves a defined purpose. Here, the purpose is essentially that AI is developed for the long-term benefit of humanity. The trustees manage voting shares in the company. Their right to appoint board members derives from these shares.
Their influence grows step by step. Initially, the Trust was only allowed to fill a portion of the seats; over time, more are added until it holds the majority. These steps are tied to fixed dates and to how much money the company has raised. The Trust itself selects new trustees; investors have no veto right in this, though special rules exist for very large shareholders.
It’s important to distinguish this from a foundation: the Trust does not own the company and receives no distributions. It only holds voting rights. You can think of it as a keyholder who decides on the composition of the oversight body but has no stake in the cash itself.
Why the term appears in business news
You won’t encounter the Trust directly in everyday life. In news about AI companies, however, it comes up regularly. Whenever Anthropic raises new capital, say from Amazon or Google, the question arises of whether the backers gain any say. The answer usually points to the Trust: capital, yes, but control over the board only to a limited extent.
The term also stands in for a larger debate. Should companies building extremely powerful technology be ordinary corporations? Or are built-in brakes needed? Comparable special constructions exist at OpenAI and at newspapers like the Guardian, which is owned by a foundation. If you read an article about corporate structure, governance, or control at AI companies, the Long-Term Benefit Trust is one of the standard examples.