Anthropic plans to allocate 5% of its shares to a Public Benefit Fund after a possible stock market debut as early as November, subject to shareholder approval. The announcement comes as the company prepares to sell investors on its commercial prospects and social purpose.

The fund would operate as a wholly owned limited liability company rather than a tax-exempt charity, with potential uses including nonprofit donations, direct distributions, and programs addressing AI’s effects on workers. Anthropic says the equity will not finance lobbying or political donations anywhere.

Its board could contribute an additional two percentage points for every 50% increase in company value above an undisclosed baseline, though those additions would remain discretionary. Anthropic has not announced a ceiling on contributions or detailed the fund’s governance.

Critics, including Palmer Luckey and David Sacks, questioned the initiative’s public value. Elon Musk suggested distributing shares directly to the public instead. Anthropic has promised further disclosures in the coming months.

But for now, the proposal specifies an initial equity commitment while leaving decisions about recipients, oversight, and future contributions largely up to the company to explain later.

Allocating shares to a fund does not give the public ownership or a say in how the money is spent; Anthropic would retain control through its subsidiary. What the proposed fund would do is concentrate decisions regarding the public benefit inside the company, whose technology is slated to produce both the wealth and the disruption. Such an arrangement could finance useful work while also allowing management to choose which consequences deserve attention and which organizations receive support.

The timing of the announcement points to a commercial consideration, as a substantial equity pledge tends to improve a company’s reputation before an offering, despite the fact that spending and measurable benefits are likely to come much later. Prospective investors are being shown a commitment whose eventual cost and operation remain partly discretionary.

Anthropic’s ban on political donations and lobbying addresses one concern, though it may not be sincere. It also does not explain how the fund will select causes or judge results.

If AI changes your job, a corporate fund doesn't automatically entitle you to compensation or retraining, because assistance would depend on programs that haven't been designed yet. Families making career decisions cannot treat a share pledge as a financial safety net. The same applies to communities facing economic disruption.

While a grant might help a local organization respond, occasional support cannot replace dependable earnings. The fund’s value on paper will matter little if assistance is not reaching people facing immediate losses.

Anthropic’s commercial success would at least ostensibly give its executives greater latitude to decide which social problems receive funding and which remedies deserve support. However, people affected by its technology would have no corresponding assurance of influence over those choices. The fund could distribute wealth widely while leaving authority concentrated.

Such is the limitation of presenting corporate philanthropy as a response to technological upheaval. Generosity remains discretionary even when the consequences for workers are not.

Anthropic may become a substantial benefactor, but the public has reason to question an arrangement in which the company helps reshape working life and then decides which consequences merit relief.