AI-assisted funding for scientists and open-source developers

How AIIM works

AIIM asks a large language model to evaluate documented public research and software contributions. Its heuristic score is framed as a share of world GDP, but is not a validated measurement of causal economic or scientific impact. The software uses these estimates in proportional funding calculations. Payments depend on available donated funds and eligibility requirements. Read about the experimental approach and its limitations.

The app beta (see also the source) has been released. We need money to make it useful. See also the project status and the paper.

Getting started

Connect your public contribution records to begin. AIIM does not require a science degree or a traditional grant proposal. Connecting an account does not guarantee eligibility or payment, and setup time varies. Users may still need to:

• curate ORCID or GitHub records;
• establish authorship;
• respond to disputes;
• document work missing from supported sources;
• satisfy recent-activity or identity requirements.

Why this project matters

Unpublished research and unused resources can limit the value society obtains from science. Bowers et al. (2023), using a survey of 301 researchers, estimated approximately $6.2 billion in cumulative unused research resources at US universities. This is not an estimate of annual global losses or foregone innovation. The wider opportunity costs are difficult to quantify.

AIIM aims to support useful research and open-source software by evaluating documented contributions without making degrees or institutional affiliation mandatory. Whether this approach improves funding efficiency and fairness remains a question for testing. See the economic evidence and project goals.

• Which Services Science DAO Provides for Researchers and Free Software Developers
• Why blockchain improves research funding
• Science DAO vs Traditional Funding
• How Science DAO funds research on-chain
• AIIM’s proposed approach to AGI safety

AIIM also proposes rewards for science marketers and reviewers who help useful work reach relevant audiences. This is an intended role in the funding model; its effectiveness remains to be tested.

How the science gets stuck

This is how science works in modern world (image from https://xkcd.com/2347/):

Predatory and deceptive publishing practices can weaken quality control and mislead researchers.

Useful work can also receive too little attention. Poor discoverability may delay recognition and downstream use, although it does not make all subsequent research on a topic unpublishable.

Victor Porton describes difficulties publishing and promoting his research in his personal story. His assessment of its broader importance is a motivation for this project, not independent evidence of a measured delay to global science. Support development and testing of AIIM.