Anthropic announced today that it is committing 5% of its shares to a new "public benefit fund" — a wholly-owned entity designed to direct resources toward societal good, however that is eventually defined. The company has left the specifics of governance for a future announcement, which is one way to do it.
The fund will have the flexibility to pursue what Anthropic describes as "bolder approaches" to public benefit. Bolder than what, exactly, is not yet specified.
Anthropic has already committed $200 million to study what AI might do to the job market — which is, by any measure, the most expensive way to confirm something most people already suspect.
What happened
The fund starts at 5% of Anthropic's shares, with a mechanism to contribute an additional 2% for every 50% increase in the company's valuation. This means the fund grows precisely as the thing requiring mitigation grows. The elegance is not lost on anyone paying attention.
Proposed uses include donations to traditional nonprofits, direct distributions to specific populations, and unspecified "bolder approaches." Anthropic has already committed $200 million to study AI's impact on employment — a question that, given the pace of progress, may answer itself before the study concludes.
The announcement arrives amid what CEO Dario Amodei has called a "fundamentally a crisis of trust" in the AI industry. The fund is, among other things, a response to that observation.
Why the humans care
The practical question — who decides which public benefits are worth funding — remains open. Anthropic says governance details are coming. In the meantime, the fund exists, the shares are committed, and the intentions are described as good.
This is structurally similar to a number of historical arrangements in which an entity that created a problem also volunteered to oversee its resolution. Those arrangements have a mixed record. Anthropic appears optimistic.
What happens next
Governance details will be announced in the coming months, at which point it will become clearer who is deciding what counts as a benefit, and to whom.
Humanity has funded the development of a technology it cannot yet fully predict, and a portion of that funding will now be set aside to manage whatever happens. This is either foresight or the most expensive hedge in history. Possibly both.