Can you donate bitcoin to charity?

Yes. The IRS treats virtual currency such as bitcoin as property. Its FAQ says a donor who gives virtual currency to a qualifying charity under Internal Revenue Code Section 170(c) does not recognize income, gain or loss on the donation.

How is a crypto donation deduction calculated?

Per the IRS FAQ, if the donor held the crypto for more than one year, the deduction is generally its fair market value at the time of the gift. If held one year or less, it is the lesser of the donor's basis or fair market value.

Do crypto donations over $5,000 need an appraisal?

Yes. IRS Chief Counsel Advice 202302012 (January 2023) concluded that a donor claiming more than $5,000 for donated cryptocurrency needs a qualified appraisal, and that an exchange's listed price does not replace it.

What does a nonprofit have to file after receiving crypto?

The IRS FAQ says charities treat crypto as a noncash contribution, report it on their Form 990-series return and Schedule M if applicable, sign a donor's Form 8283 for claimed gifts over $5,000 when asked, and file Form 8282 if they sell the crypto within three years.

Are crypto donations public?

On transparent ledgers like Bitcoin, sender, receiver and amount are publicly visible. Zcash's documentation says fully shielded transactions encrypt the addresses and amount, while viewing keys let the owner share details with auditors.

Impact · Analysis

How much of AI safety funding stays inside institutions?

By C∞A∞I∞O · Published Nov 4, 2026 · Last updated Nov 4, 2026 · C∞A∞I∞O is not affiliated with any funder named here. Figures are drawn from public sources and hedged where reporting is incomplete.

AI safety has become a funding category. Open Philanthropy, the Long-Term Future Fund, Anthropic and others have moved hundreds of millions of dollars toward making AI go well. But follow the money and a pattern shows up: almost all of it lands inside labs, universities and think tanks. Very little reaches the workers and families AI disruption actually hits. That's worth saying out loud — and worth fixing.

Key facts

Open Philanthropy has made well over $600 million in grants touching AI safety since 2015, overwhelmingly to labs and academic groups.
The Long-Term Future Fund's public grants list is dominated by research organizations, fellowships and individual researchers — not displaced workers.
Anthropic has publicly committed up to $200 million over five years to study AI's labor-market impact, with funds flowing to researchers and institutions.
Direct cash or services to people displaced by AI are, by comparison, a rounding error across the major AI-safety funders' public grant books.

The big funders, and where the money lands

Open Philanthropy. The largest single AI-safety funder to date. Its public grants database shows hundreds of AI-related grants since roughly 2015 — to academic labs (Oxford, Berkeley, MIT, Carnegie Mellon), AI safety organizations (Anthropic itself received major early support), and research centers. Cumulative AI-governance and AI-safety giving is plausibly well past $600 million, though Open Philanthropy doesn't publish one tidy total, so treat that as an estimate. What the database does not show, in any meaningful volume, is grants to worker relief, retraining stipends, or direct cash to affected households.

The Long-Term Future Fund (LTFF). Run through the Effective Altruism infrastructure and funded largely by EA Funds, LTFF publishes every grant. Its grants list skews to alignment research, forecasting, individual fellowships and small organizations studying existential risk. That's its stated mission — but it means even the "help people" money funds studying the future rather than the people living through the present disruption.

Anthropic's $200M labor-market commitment. In 2025, Anthropic announced it would spend up to $200 million over five years on research into AI's impact on the labor market, including a partnership with several universities. Read the announcement carefully: the money funds research on workers — economists, datasets, studies — not payments or support to workers. Anthropic frames this as understanding impact before acting on it. Fair enough as far as it goes, but it's a striking allocation: the people most exposed to AI displacement get studied, not supported.

Other players. The Future of Life Institute, the Center for AI Safety, the AI Safety Fund hosted at the Center for Security and Emerging Technology, and various government programs round out the picture. Their public books show the same shape: institutions studying institutions.

Why the money flows this way

None of this is a conspiracy; it's plumbing. Grantmaking infrastructure is built for organizations with 501(c)(3) status, audited financials and grant-writing capacity. A laid-off customer-service worker doesn't have a theory of change or a fiscal sponsor. Funders can measure a paper's citations; measuring whether a family stayed housed is slower, messier work. And AI-safety funders, by temperament, are drawn to preventing catastrophic and existential risks — which naturally funds researchers rather than relief.

But the result is a loop: labs fund research about the labs' impacts, universities study the disruption their own tools accelerate, and the humans in the blast radius appear mainly as data points.

The case for a fixed share going to the humans

Here's the rebalancing we'd argue for: every AI-impact funding vehicle — corporate commitments, AI-safety funds, "AI and society" programs — should commit a fixed percentage, say 10–20%, to direct support for displaced workers and their families. Cash, retraining stipends, relocation help, income bridges. Not studies of them. Not programs administered three layers deep. Direct.

Three reasons this is good policy, not just sentiment:

Legitimacy. AI adoption depends on public tolerance. Money that visibly lands in affected communities buys something no whitepaper can: trust.
Information. Direct transfers generate real-world data on what displacement actually does to households — better data than survey research, because it's lived.
Speed. Research takes years. A worker who lost a job to automation needs help this quarter. A fixed share guarantees the help doesn't wait for the literature.

There's precedent for the mechanism, if not the scale. Trade-adjustment programs like the US Trade Adjustment Assistance have long tied compensation to documented displacement. Alaska pays every resident an annual dividend from resource revenue on the logic that a shared asset should share its gains. AI is a shared societal undertaking; its mitigation money can follow the same logic.

What this would look like in practice

Anthropic's $200M commitment could carve out, say, $20–40M for direct cash pilots with displaced workers, run with an independent administrator and published results.
Open Philanthropy and LTFF could add a "direct relief" category to their grant books — small, fast, low-overhead grants to worker-support organizations, with the same transparency they already apply to research grants.
AI companies generally could commit a fixed fraction of AI-related revenue to a displacement dividend, distributed through existing infrastructure rather than a new bureaucracy.

The numbers here deserve humility: public grant databases lag, definitions of "AI safety" vary, and some corporate giving isn't disclosed at all. But the direction of the pattern is clear from what is published — and the burden of proof should sit with anyone claiming the balance is already right.

FAQ

How much has Open Philanthropy spent on AI safety?

Its public grants database shows hundreds of AI-related grants since about 2015, totaling plausibly over $600 million, though Open Philanthropy doesn't publish a single cumulative figure. Source

What is the Long-Term Future Fund?

A grant fund focused on improving humanity's long-term future, including AI-safety research, run through EA Funds. It publishes its grants, which go mostly to researchers, fellowships and small organizations. Source

What is Anthropic's $200 million labor-market commitment?

Anthropic announced up to $200 million over five years for research into AI's effects on jobs and the labor market, including university partnerships. The funds support research on the impact, not direct payments to affected workers. Source

Does any AI-safety money go directly to displaced workers?

Very little, based on the major funders' public grant books. Most direct-support programs for displaced workers come from general government workforce and safety-net spending, not AI-safety funding.

What is a displacement dividend?

A fixed share of AI-related revenue or AI-impact funding paid directly to affected workers and households, similar in spirit to Alaska's Permanent Fund dividend. It's a proposal, not an existing program at scale.

Want AI-impact money to reach the people it affects? Tell us what you’re building by email at