A decade ago, Ryan Carrier was a failed hedge fund manager watching as AI systems began to spin out of control. Facebook’s algorithms shook US elections, a Microsoft chatbot said the Holocaust was “made up,” and Tesla’s Autopilot killed its first driver. “There was no governance, oversight, or accountability,” Carrier says. To him, the AI-filled future awaiting his children looked bleak.
Carrier went on that year to found ForHumanity, a nonprofit organization developing tools for auditing AI systems. But it has raised just hundreds of thousands of dollars since 2016 and is still a minor player in the industry. But that could soon change, because ForHumanity is now among the many nonprofits around the world trying to rein in AI, get animals out of cages, eliminate poverty and sickness, and spread democracy that are eager to claim a slice of what’s widely anticipated to be the largest wave of philanthropy in decades.
Two nearly trillion-dollar AI companies, ChatGPT developer OpenAI and Claude creator Anthropic, are expected to go public soon, making hundreds of current and former employees ultrawealthy. Some of them are aligned with a philosophy known as effective altruism that encourages making impactful donations sooner rather than later.
Anthropic’s seven founders have pledged to donate 80 percent of their wealth, and the company has agreed to chip in one or three shares for every one its employees commit to giving, depending on when they joined and up to a certain limit.
Rough estimates by a tech industry insider suggest that Anthropic’s IPO, which could happen in September, may result in $15 billion a year in additional philanthropic giving alone. That would be enough to boost total US giving by about 2.5 percent annually—the same as adding four Bill Gates’s, one of the world’s biggest donors. Anthropic declined to comment on the total amount its employees have set aside to donate and the organizations that may benefit.
All that money is far from guaranteed. The IPOs may get delayed or go poorly and leave employees clinging onto their wealth. Industry observers are also concerned that a paralyzing number of charitable options and natural fickleness could prompt workers to keep more for themselves than anticipated.
But competition for would-be donors’ attention is already fierce. Jack Lewars, a consultant who advised 13 ultrarich tech and finance workers on their charitable giving last year, says he’s heard that employees at the AI labs are receiving as many as 20 unsolicited emails a week from groups seeking donations.
WIRED spoke with 18 nonprofits and approached dozens more to ask how they’re preparing for the potential philanthropy windfall. None of them admitted to sending cold pitches—a tactic that Lewars on his new blog, The Funding Anthropalypse, wrote “has next to no chance of working.” Instead, organizations say they are ramping up hiring, training, marketing, and use of automation to position themselves to attract massive sums of funding and quickly put it to use. One job posting at an education nonprofit even calls out building up relationships at Anthropic specifically as a priority.
“Everybody’s going to go after these funds,” says Christine Peterson, cofounder of the grantmaking group Foresight Institute, which claims to have funding from Anthropic employees. “It’s going to be a wild ride.”
The Inside Track #
Like many nonprofit leaders who spoke with WIRED, ForHumanity’s Carrier says he is focusing more on the work itself than on fundraising. But he recognizes this may be a moment to shift, and he has begun wondering how to get into IPO soirees in San Francisco. “I just have to get in that room,” Carrier says.
Bo Young Lee, CEO of the nonprofit AI4ALL, says she is attending more events, putting out more research, and asking board members, such as AI scientist and entrepreneur Fei-Fei Li, to introduce her to employees at the AI labs. Her organization trains young adults across the US how to develop their own AI models, with the goal of diversifying the tech workforce. Lee says she is setting “ambitious” fundraising goals because she’s confident the money will come, though introductory meetings haven’t materialized yet.
Buck Shlegeris is the CEO of Redwood Research, a Berkeley, California-based nonprofit that is part of a growing set of small organizations pursuing what’s broadly defined as AI safety. From his perspective, targeting individual donors isn’t an optimal strategy.
Redwood has received millions of dollars from grantmaking groups such as Coefficient Giving and Survival and Flourishing Fund that pool together donations from individuals and subscribe to the effective altruism movement. Shlegeris believes that new money will flow to these intermediaries and then continue to trickle down to Redwood and others.
He says he wants to accelerate training staffers on becoming managers with the expectation that big sums of money will enable teams to grow and take on “crazy expensive projects,” like automating safety research and training Redwood’s own models. The nonprofit’s overall goal is to minimize the risk that AI could somehow lead to human extinction, which Shlegeris says he fears has a “really strong chance” of happening.
A subset of AI safety research is centered on preventing the technology from being used in the creation of bioweapons. Venture capitalist Geoff Ralston recently helped author an action plan that calls for raising $2.5 billion over the next five years to address AI biosecurity. He plans to solicit donations from those who stand to benefit from the IPO windfall. “The folks at frontier labs understand the threat vectors created by AI better than anyone,” Ralston says.
Several grantmakers influential in effective altruism circles are preparing for the giving wave by helping newer nonprofits level up their administration and bookkeeping. The goal is to ensure more organizations are ready to spend the expected influx of donations. “We’re trying to build the port before the ship arrives,” says Stien van der Ploeg, executive director at Animal Charity Evaluators, which over the past year helped direct about $15 million to nonprofits trying to reduce what they view as the worst forms of farm cruelty.
Similarly, Coefficient is making a push to support up-and-comers. This month, its largest donors, Facebook cofounder Dustin Moskovitz and his wife Cari Tuna, committed $1 billion to global health projects. It’s a “one-off surge” nearly six times bigger than initially planned but made with the intention of creating “scalable opportunities” that can “effectively absorb much higher amounts of future giving,” according to Coefficient.
GiveDirectly, another organization popular among effective altruists, says it discretely raised a round of funding from donors to support its preparations for the giving wave. The nonprofit, which transfers unconditional cash to people in poverty or crisis, is using the money to bring on more engineers to automate its finance and HR systems. It’s also forging partnerships to deploy money faster during natural disasters and has begun developing a plan for “a global AI wealth dividend” to fund people in extreme poverty. GiveDirectly CEO Nick Allardice says “despite the uncertainty” around the IPOs, it’s a moment “worth taking very seriously.”
More Cautious #
Some organizations worry they'll be left behind because they're removed from the San Francisco AI community or work on issues such as child safety or political disinformation, which may not overlap with the assumed priorities of potential donors.
One broadly shared anxiety is that the sheer amount of money could leave some causes with overflowing coffers, while other urgent issues may attract little funding. For instance, groups addressing the existential risks AI poses to humanity—a priority for effective altruists—are widely predicted to receive far more backing than those trying to improve human rights by, say, fighting mass surveillance or online harms.
That possibility has been keeping Marlena Wisniak up at night. She oversees digital strategy at the European Center for Not-for-Profit Law (ECNL), which works on AI policy and research, and has been trying to raise the profile of organizations that may be overlooked, especially in the global south. This month, she scored a victory when a friend who works at Anthropic donated $100,000 to one group that fits that description. Wisniak is now trying to convince contacts at OpenAI and Anthropic to share her list of human rights and social justice organizations worth supporting with their colleagues. She is also encouraging those nonprofits to frame their work using terms like “theory of change” and “evidence-backed” that might better resonate with effective altruists.
A few organizations are deliberately sitting out the moment. Earlier this year, Model Evaluation and Threat Research, which evaluates OpenAI and Anthropic models, decided against soliciting funding from employees at those companies because it could jeopardize the nonprofit’s independence. Other groups are concerned about receiving funding from sources that would link them to effective altruism and scare away partners or other donors because the movement has been criticized as insular and misguided, according to a person familiar with the thinking of the nonprofits who sought anonymity to discuss a sensitive issue. (A communications adviser to several organizations aligned with effective altruism but not authorized to speak on their behalf says the movement “has continued to grow its funding, talent, and impact, and the increasing willingness of major funders to work with EA-aligned groups reflects that these purported reputational fears are overblown.”)
Some veterans of the nonprofit industry are urging general caution. They don't want groups to neglect their core work by contorting their projects to fit a mold appealing to the new money. It's also not lost on fundraisers that the wealth is a byproduct of building AI tools that, in some cases, are worsening the problems that nonprofits are tackling. “The risk today is this industrialized wealth from these IPOs may not serve human good in hindsight,” AI4ALL’s Lee says. “We have to avoid the allure of easy money simply to appease the prioritization of the wealthy.”
One effect of the anticipated funding surge is already emerging: As philanthropic funding concentrates in a handful of fields with limited talent pools, salaries are rising. This month, Resolution, an AI safety nonprofit, announced a $160 million grant from Coefficient, the donor’s largest award of its kind. The funding combined with the “enormous influx of philanthropic capital” following the AI IPOs will allow Resolution to pay “well above nonprofit and academic norms,” the organization said in a blog post.