Value Compass

Raising $45,000active

A values and concentration map that exposes the values of AI models and tools makers, and those behind the makers (funders, jurisdictions etc) to the public (individuals and organizations) so they can vote with their choices.

Project Details

Value Compass does a few things:

It maps the connections between and behind AI powers that are not as obvious: who's funding which companies/projects, what jurisdictions they fall under, who are these companies and funders connected to, etc. It makes the opaque power connection map transparent.
It tracks the values (both shown and stated) by the AI makers and backers: currently on five axis: Transparency, Culture / ESG, Labour & supply-chain integrity, Wealth dispersion, Public wealth-sharing, all with data and prove. The axis is v0 and to be refined.
It exposes the power map and values to the public in a way that's easy to use so individuals and organizations can make choices based on this information. Current V0 of consumer facing interface is here, with plans to add a lot more makers and funders (it is only genuinely useful with more granular makers and funders information and better usability), as well as making this information available via API.
I'm working on this independently but may seek additional help on research, design, or maintenance if funded.
The plan is for a 6-month build to a functional public product

Theory of Impact

Power concentration is one of the most worrying x-risk from AI, largely because concentrated powers won't be accountable to the public--the powers would not need to optimize for the welfare of the public, and thus they likely won't.
This project does two things: it exposes the hidden connections of concentrating powers to the public, and it makes clear what values these powers hold through their words and actions.
That allows the public to use their power of votes and purchasing powers, which is one of the last remaining powers they have for a short window to push for this set of outcomes: ideally, the concentration of power will slow down as the public has the information to support the pro-social, more independent makers of AI; and worse case scenario, even if power concentration still continues, the public may be able to select for the more pro-social powers to win over the other ones.

People

Yuyu Shen

Team Member

Grants Received– no grants recorded

Add grant

Funding Asks

grantmaking.ai Launch Round

Applied

Minimum

$45,000

Ideal

$205,000

How the money will be spent

Numbers are converted to British pounds:

Minimum: 6 months project, founder/lead salary £30K; compute, tooling, and tech ops (LLM inference for assessment and the user-facing interface, plus API hosting), ~£3K; and reach/partnerships to reach high-leverage users, ~£0.6K.
Ideal: 12 months project, founder/lead at 70% time, £70K (70% of a £100K yearly full-time rate); a part-time research contractor (~5–6 days/month), ~£24K, validating and maintaining the power-and-values data layer; a part-time developer (~4–6 days/month), ~£30K, building the API, infrastructure, and interface so founder time stays on methodology; compute, tooling, and tech ops (LLM inference for assessment and the user-facing interface, plus API hosting), ~£14K; and reach/partnerships to reach high-leverage users, ~£8K. Includes a fiscal-sponsor fee of up to ~10%.

Discussion

ZZeming7d

This is a great project!
What is the data source tho? How credible is it?