Mapping AI

Who actually shapes AI policy in the U.S.? We mapped 1,812 entities: 745 people, 918 organizations, 2,925 relationships. Frontier Labs, AI Safety orgs, Think Tanks, Government, VCs, and more. mapping-ai.org

shout out to @JustenMichel for his AI tribes post that we referenced in first trying to make sense of this space! substack.com/home/post/p-18… would love any feedback you have :) or a collab? 👀

Hey @ohlennart - thank you so much for the feedback! This thread has more context on why we built this and how we think about the data and its current gaps: x.com/mapping_ai/sta… Our approach is to capture what's being discussed both internally and externally about entities in the AI policy space. Many think tanks and advocacy orgs are officially nonpartisan but are widely perceived as leaning a certain way on regulation. We think capturing that perception alongside stated positions is valuable, and we note on the site that inferred stances don't claim to represent official positions (while also respecting orgs' corrections). We’re open to suggestions on methodology here and how to more accurately capture these nuances! We also acknowledge there are errors, and we've already corrected several entries today thanks to direct outreach from organizations, including removing stances entirely where orgs don't want to take public positions. We'd rather ship something correctable than wait for perfection, but we aim to update promptly: both when groups reach out and when they self-contribute their data. The data and code are open-sourced specifically so the community can help improve it. We want to get entries right and will prioritize corrections as they come in: mapping-ai.org/contribute, email info@mapping-ai.org, or our Discord discord.gg/mappingai. If there's anything specific you'd like corrected, let us know :)

C'mon this seems to be vibe-coded without any review. - Lots of views attributed to individuals despite it just them stating a viewpoint in an interview. - Organizations with viewpoints despite these orgs explicitly not taking positions (and it's just individuals saying that). - People being affiliated with orgs they're not affiliated with. - What are even partner organizations? At least pay for Opus tokens and prompt better.

Who actually shapes AI policy in the U.S.? We mapped 1,812 entities: 745 people, 918 organizations, 2,925 relationships. Frontier Labs, AI Safety orgs, Think Tanks, Government, VCs, and more. mapping-ai.org

@davidshor @NateSilver538 @GaryMarcus @Miles_Brundage @timhwang We mapped 373 AI leaders on regulation stance vs AGI timeline. The investor cluster sits firmly in the bottom-left: short timelines, minimal regulation. What’s your read?

Hi @PeterSlattery1, thanks for taking the time to review + flagging this! We’ve updated MIT FutureTech's AGI timeline to Unknown - let us know if you have any other corrections. We kept the LinkedIn source linked with inference set to "low" since it’s related but from a personal post. In general, our data comes from public info, user submissions via our contribute form (mapping-ai.org/contribute), and LLM-assisted research. We note on the site that listed and inferred stances are not necessarily official positions - the goal is to capture what's being discussed publicly (we see internal and external perceptions on stances as signal, not noise), while also respecting orgs' officially stated stances and corrections. We aim to update promptly, both when groups reach out and when they self-contribute their data.

How did you validate this data to ensure accuracy? Did you check any of these with the organizations or individuals being mapped? I ask as I am concerned about the quality of the evidence I have seen cited in this. For instance, this excerpt from my LinkedIn post is presented as the evidence for saying that the MIT FutureTech lab has a 5-10 year AGI timeline. How is the below text from my post evidence for a lab having a specific timeline? -- AGI Timeline 5-10 year timeline medium "We had a very interesting discussion exploring our different perspectives on AI timelines. We didn't have much time to get into details during the short discussion, but the high level summary was that many people in FutureTech were more pessimistic about a few aspects of the scenario. They expect that AI will be slower and less effective at rapidly speeding up AI research and have slower and lower adoption in key sectors and across the economy as a whole." ---

These are just a few insights from the map. The tool is only as strong as the data, so please help us keep adding (mapping-ai.org/contribute). Explore the full project here: x.com/mapping_ai/sta…

Explore the interactive version by clicking any dot to see the full profile. mapping-ai.org/map (switch to Plot view) #outliers" target="_blank" rel="nofollow noopener">mapping-ai.org/insights#outli… (or scroll to Outliers)

We plotted 373 AI leaders on two axes… X: How much regulation do they want? Y: When do they think AGI arrives? …and some striking outliers emerged.

@JFinnyc @Zai_org Absolutely. We started with the US AI policy ecosystem out of familiarity but we hope to scale this tool into an international mapping project. If you're interested in collaborating or know of potential collaborators, please reach out over DMs or info@mapping-ai.org.

someone should make a proper mapping of china AI. it would be the perfect visual to show its bào tuán culture. not just who is building what, but the layers underneath: 1️⃣ city alignment which frontier lab is tied to which city / municipal government @Zai_org -> beijing @StepFun_ai -> shanghai @deepseek_ai -> hangzhou 2️⃣ lineage networks + tight in-group clustering which university, company, and inner circle the researchers come from tsinghua, beihang, ex @Megvii, ex @SenseTime_AI, etc. a lot of china AI only really starts making sense once you see that each lab is shaped not just by product direction, but by city backing, school lineage, and ex-company networks.

@ItsCuthulhu Thanks for sharing Jake! Great work tracking forecasts, I'm including this in our "Adjacent Tools & Resources" section. Always open to collaboration to make these tools as comprehensive as possible for the community. Feel free to reach out over DMs or email: info@mapping-ai.org

@mapping_ai I actually have many similar labs related to AGI on my website. One of them is about when people said AGI would happens, and maps 250+ sources: jakecuth.com/work/agi-forec…

Everyone keeps saying "AGI is coming." But when you actually map what people mean by AGI, you find completely different definitions. We built a tool to visualize this. Here's what the data shows:

@DrTechlash Incredible, thanks for sharing @DrTechlash. Funding is such an important flow to track, and that’s the next layer we’re starting to build #funding" target="_blank" rel="nofollow noopener">mapping-ai.org/insights#fundi…. DMing you for sources on this Coefficient Giving funding, excited to incorporate your rich data.

Hey @mapping_ai team, I made this for you. Updated numbers you can use.

If this is a valuable resource, we also want to figure out how to sustain it long term (we’re just a couple of 24 year olds with laptops). If you have ideas on how to scale open civic tech and want to support: email us at info@mapping-ai.org!

There are MANY many tracker tools (e.g., meta tracker tracker ai-policy-tracker-tracker.vercel.app @Daniel_Kalish_) we’re documenting in “Adjacent Tools.” If you're building these, we’d love to collaborate rather than duplicate work. We’ve open-sourced our code base: github.com/MappingAI

Getting a lot of questions about Mapping AI so here's some context. We are concerned and hopeful about AI, but as outsiders from climate, planning, space, housing, and organizing, we didn't know where to begin. So we mapped it out 🗺️and open sourced it all:

@neil_chilson 👀 whoops @neil_chilson, you're in there now 🫡! (and if anything's amiss, let us know via mapping-ai.org/contribute or email us at info@mapping-ai.org)

Yikes, Fran Drescher is in here but I’m not. Gotta up my game.