Articles:
There should be more AI safety organisations
Why does the AI Safety Community need help founding projects?
Alignment can be the ‘clean energy’ of AI
AI Tools for Existential Security
What makes an AI Startup “Net Positive” for Safety
Incubation Programs:
Def/acc at Entrepreneur First—this is a new program, which focuses on “defensive” tech, which will produce a wide variety of startups, some of which may help with reducing x-risk, but others may include AI applications like autonomous weapons or diagnostic medical software, which, while they generally won’t increase x-risk, may not reduce it either.
Seldon Accelerator: “We take on AGI risk with optimism”
Constellation Incubator: Supports pre-idea entrepreneurs and early-stage organizations
AE Studio has expressed interest in helping seed neglected approaches, incl. ideas that take the form of business ideas
Fifty Years: “We help AI researchers build startups for a safe and aligned future”
AIS Friendly General Program:
Founding to Give: General EA incubator
Entrepreneur First—They run def/acc, so this could be a decent alternative
Halycon Futures: “We’re an entrepreneurial nonprofit and VC fund dedicated to making AI safe, secure and good for humanity.” They do new project incubation, grants, investments, ect.
Brains Accelerator: “Brains is an accelerator that helps talented scientists and technologists execute on ambitious research visions that are beyond the scope of individual academic labs, startups, or large companies. These visions range from upending the way we deal with carbon to how we understand chronic disease or observe the universe.”
Communities:
Entrepreneurship Channel EA Anyway
AI Safety Founders: Website (with links to resources), Discord, LinkedIn, collaboration request
VC:
AI Safety Seed Funding Network
Lionheart Ventures—“We invest in the wise development of transformative technologies.”
Juniper Ventures—Invests in companies “working to make AI safe, secure and beneficial for humanity”. Published the AI Assurance Tech Report
Polaris Ventures—“We support projects and people aiming to build a future guided by wisdom and compassion for all”
Mythos Ventures—“an early-stage venture capital firm investing in prosocial technologies for the transformative AI era.”
Metaplanet—Founded by Jaan Tallinn. “We invest in deep tech that benefits humanity in the long run. We also make grants to fund projects that don’t fit the venture model… We also love projects that reduce existential risks from AI and other advanced technologies. We tend to skip well-known immediate risks and remedies that get ample attention and investment”
Anthology Fund—Backed by Anthropic. One of the five key areas is “trust and safety tooling”
Menlo Ventures—Lead the GoodFire round (article)
Babuschkin Ventures (contact): “Supports AI safety research and backs startups in AI and agentic systems that advance humanity and unlock the mysteries of our universe”
Safe AI Fund—“an early-stage venture fund dedicated to supporting startups developing tools to enhance AI safety, security, and responsible deployment. The fund provides both financial investment and mentorship”
Organisational Support:
Ashgro—fiscal sponsorship
Rethink Priorities Special Projects—provides fiscal sponsorship or incubation
Other:
Hackathon for AI Safety Startups—Ran once by Apart Research, may run again
Constellation Residency—Year-long position
Nonlinear—Free coaching for people who are running an AI safety startup or who are considering starting one