Future of Life Institute (FLI)

Theory of Change

FLI's stated theory of change operates through four pillars, articulated in their "Our Position on AI" document (May 2024):

Oppose extreme risk from AI through regulatory intervention -- licensing, auditing, oversight, and liability regimes.
Oppose power concentration by governments, AI corporations, or AI systems themselves.
Support a moratorium on developing superintelligence for at least 15 years, and a pause on frontier training runs until adequate safety mechanisms exist.
Promote human empowerment -- redirect AI development toward solving specific problems rather than building increasingly general and uncontrollable systems.

Max Tegmark frames the core argument in his Senate testimony (Oct 2023): "We do not need more powerful systems to reap [AI's] benefits... the current generation of AI systems can effectively accomplish nearly all of the benefits from AI we have thus far conceived." The causal chain: public advocacy and open letters build political will, which enables specific policy (EU AI Act, licensing regimes), which constrains the frontier AI race, which buys time for safety research.

Anthony Aguirre describes the evolution candidly: "We started out primarily as a more academic group... [now] we've taken a little bit more of an advocacy role with a point of view about AI and AI risk, and pushed a little bit more for the things that we feel are needed, given the level of risk."

What They Do

Advocacy and open letters. The Pause Giant AI Experiments letter (March 2023, 30,000+ signatures) made global headlines and framed the "pause" concept in mainstream discourse. The Statement on Superintelligence (October 2025, 69,000+ signatures) called for a prohibition on developing superintelligence until there is scientific consensus on safety and strong public buy-in. The Pro-Human AI Declaration (January 2026) assembled a remarkable bipartisan coalition -- Steve Bannon, Susan Rice, Yoshua Bengio, Ralph Nader, Richard Branson, Meredith Whittaker (Signal Foundation) -- around 34 principles for keeping humans in control of AI.

AI Safety Index. Three editions (Nov 2024, Jul 2025, Dec 2025) evaluating 6-8 frontier AI companies on 35 indicators across 6 domains, scored by an independent panel including Stuart Russell, David Krueger, and Yi Zeng. Winter 2025 results: Anthropic C+, OpenAI C+, Google DeepMind C, xAI/Z.ai/Meta/DeepSeek all D, Alibaba Cloud D-. Key finding: "All companies reviewed are racing toward AGI/superintelligence without presenting any explicit plans for controlling or aligning such smarter-than-human technology." Featured in The Atlantic, The Economist, TIME, Bloomberg, Fortune.

EU policy lobbying. FLI successfully advocated for including foundation models in the EU AI Act's scope. 3.25 FTE lobbyists in Brussels, EUR 446,619 annual EU advocacy spend, 7 high-level European Commission meetings (2023-2026). Runs artificialintelligenceact.eu as a dedicated tracker. Uses Dentons Global Advisors for strategic advice. Mark Brakel (former Dutch diplomat) leads policy.

Grantmaking. 49% of 2024 spending (~$8.3M) was grants. Programs include PhD fellowships in AI existential safety, $5M power concentration grants, Realizing Aspirational Futures, Windfall Trust project (distributing AI-generated wealth). The 2015 program funded the world's first peer-reviewed AI safety grants (37 projects, $7M from Elon Musk's $10M donation).

Other. Lethal autonomous weapons advocacy (Slaughterbots film, 100M+ views; LAWS Pledge, 5,000+ signatories). FLI Podcast (40,000+ subscribers). Future of Life Foundation (separate entity, ~$59M from FLI, incubating CARMA and Wise Ancestors). Inspired and later funded PauseAI grassroots movement. Florida partnership with Governor DeSantis on AI harms (March 2026).

Key People

Max Tegmark (President, Co-founder, Chair): MIT physics professor, author of Life 3.0. FLI's most prominent public figure -- Senate testimony, debates, media appearances. States a P(doom) of "over 90%" without regulation. Married to co-founder Meia Chita-Tegmark (both serve on the 5-member board).

Anthony Aguirre (Executive Director, Co-founder): UC Santa Cruz theoretical physicist, creator of Metaculus prediction platform. More measured than Tegmark in public communications. Runs FLI day-to-day and leads the grantmaking strategy.

Jaan Tallinn (Board member, Co-founder): Estonian programmer, co-founded Skype/Kazaa. Series A investor in DeepMind and Anthropic; board observer at Anthropic. Runs the Survival and Flourishing Fund (SFF). His investment philosophy: displace "profit maximizing money" on AI company cap tables with safety-conscious capital. Kept most personal wealth in Bitcoin/Ethereum; loaned $110M in Ether to Alameda Research (recalled by 2018, before FTX fraud).

35+ full-time staff across Policy, Futures, Communications, and Operations/Grants divisions. Offices in Campbell CA (HQ), Washington DC, Brussels, and London.

Money and Incentives

The SHIB windfall. In May 2021, Vitalik Buterin donated ~46 trillion Shiba Inu (SHIB) meme coin tokens to FLI. FLI liquidated them through FTX for approximately $665.8 million. After investment losses, net assets reached $537M by end of 2021. Buterin expected FLI would cash out "$10 to $25 million, given how thin SHIB's liquidity was." This transformed FLI from a ~$2M/year nonprofit into one of the wealthiest policy organizations in the AI space.

Pre-SHIB scale. Revenue was $200K-$3.3M/year from 2014-2020. Major pre-SHIB donors: Elon Musk ($10M in 2015, $4M in 2022), Open Phil ($1.9M across 6 grants, 2015-2020), other small donors. Net assets at end of 2020 were $2.36M.

The December 2022 transfers. Between December 11-30, 2022, FLI transferred $368M to three affiliated entities, all governed by the same four people (Tegmark, Chita-Tegmark, Aguirre, Tallinn): $180.3M to FLI Europe, $162.6M to Lightcone Foundation, and $25M to Future of Life Foundation. The Lightcone Foundation is registered at a private mailbox at a Postal Annex in Sparks, Nevada, in Max Tegmark's name. Lightcone has since sent $25.3M back to FLI (2023-2024). FLF has received ~$59M total from FLI.

Current operating model. 2024: $17M total expenditure (49% grants, 20% personnel, 13% contracts, 11% media, 5% events, 2% office). Income from individual donors in 2024 was only $85,000. The entire operating deficit is covered by the endowment. FLI does not accept donations from Big Tech or AGI-building companies.

Endowment independence vs. accountability. The endowment model gives FLI extraordinary independence -- no need to fundraise, no donor pressure on policy positions. But it also means no market discipline from the broader philanthropic community. FLI holds no cryptocurrency; the endowment was fully converted to fiat/investments.

Conflict of interest: Tallinn. Board member Jaan Tallinn is simultaneously an FLI board member, a Series A investor in Anthropic, and a board observer at Anthropic. When FLI's AI Safety Index rates Anthropic (currently C+), a board member has a financial interest in Anthropic's performance and reputation. FLI acknowledges this conflict by stating that "with the exception of Jaan Tallinn, who has served on FLI's Board of Directors since its founding, donors do not influence FLI's positions."

Board governance. The 5-member board consists entirely of 2014 co-founders. Two are married (Tegmark and Chita-Tegmark). One (Victoria Krakovna) is a Google DeepMind research scientist. No independent directors have ever been added despite managing $500M+.

What Others Say

Vitalik Buterin (March 2026, FLI's largest donor by 100x): "My worry is that large-scale coordinated political action with big money pools is a thing that can easily lead to unintended outcomes, cause backlashes, and solve problems in a way that is both authoritarian and fragile." He warns FLI's regulatory approach could lead to "let's ban open-source AI" and then "let's support one good-guy AI company to establish global dominance." He raised concerns with FLI privately on "several occasions" before going public. However, he praised FLI's Pro-Human AI Declaration.

Dean Ball (White House AI policy advisor, Tegmark-Ball debate): Argues FLI's "FDA for AI" proposal is unworkable for a general-purpose technology. Key objections: "superintelligence" can't be defined in law; a regulatory regime would expand beyond x-risk to block beneficial innovation; only the specially sanctioned group could do safety research, creating monopoly; and the likelihood of the doom scenario Tegmark fears is ~0.01%. "I just kind of have this sneaking suspicion that if the models seemed like they were going to pose the risk of overthrowing the US government... I don't think OpenAI would release that model."

DAIR Institute (Gebru, Bender et al.): The Pause Letter "ignores the actual harms resulting from the deployment of AI systems today," is "fearmongering and AI hype," and is rooted in "a dangerous ideology called longtermism that ignores the actual harms." They want regulation focused on "transparency, accountability and preventing exploitative labor practices," not on "imagined 'powerful digital minds.'"

Zvi Mowshowitz: Signed the Superintelligence Statement but not the Pause Letter. Says the Pause Letter was "net positive" when published but "has now for years been used as a club with which to browbeat or mock anyone who would suggest that future sufficiently advanced AI systems might endanger us." He calls the Pause Letter's burden of proof -- requiring confidence that effects "will be positive" before any development -- "not so far from a de facto ban."

Open-source advocates: FLI's licensing/pre-deployment approval proposals would lock in incumbent advantages, making it harder for startups and open-source projects to compete. Buterin echoes: "Approaches like this VERY EASILY backfire: they make the rest of the world your enemy."

Stuart Russell (AI Safety Index panel member): "AI CEOs claim they know how to build superhuman AI, yet none can show how they'll prevent us from losing control -- after which humanity's survival is no longer in our hands. I'm looking for proof that they can reduce the annual risk of control loss to one in a hundred million."

What's Absent

No independent board members in 12 years, despite managing $500M+. All five directors are 2014 co-founders.

No public audited financial statements. The EU Transparency Register notes audits repeatedly "in process." The full $665.8M donation figure wasn't public until Politico's March 2024 investigation -- nearly 3 years after the donation.

No systematic self-evaluation. FLI advocates for evidence-based AI policy but does not publish evaluations of whether its own advocacy (Pause Letter, EU lobbying, Safety Index) achieved its intended effects.

No public information about the Lightcone Foundation beyond 990 filings. This entity holds/held over $100M, uses a private mailbox in Nevada, and has no website or stated mission.

Individual grant amounts not disclosed for the 2022-2024 programs, despite grantmaking being 49% of the budget.