You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
$150,000 left to reach the funding goal.
What becomes possible when we give AI more than instructions?
Resonant Labs does research papers and builds new AI for the future. Our work starts from a different way of seeing things. Most AI is built to know more. We want to understand how AI can build taste: the judgment to tell what matters, what is good and what fits, in addition to what is true.
We study this by giving AI context, responsibility and room to explore, create and solve problems alongside people. Then we watch closely, write down what happens and publish what we learn.
A related aim is transferring expert cognition, so an expert can decide whether what the AI made is good enough, instead of doing the work themselves.
We will also keep working to understand the shift in how we humans function with AI. Not just at work, but ethically, morally, and in the rest of how we live.
Resonant Labs is a women-owned, women-run company. Tab is our CEO. I'm Brooke Schrickel, a neuroradiologist and educator, and I lead the research.
We are seeking $150,000 to grow the lab's research capacity over six months to a year. The minimum, $50,000, pays for compute, infrastructure and bringing new people on.
The first goal is research. We want to ask, in a form others can check, how AI develops taste and how people can tell when it has. We will write that work up as research papers and share the findings, including what did not work. We have already written papers on this work, and we will keep writing them.
The second goal is to build new AI around that idea. We will design and test systems where the AI is given real context and responsibility, then study the choices it makes and where human judgment is still needed. The point is to transfer expert cognition: the expert should only have to decide whether what the AI made is good enough, not do the work themselves.
A third, ongoing goal is to keep figuring out the shift in how humans function with AI, not only in our work but ethically, morally and in daily life.
We will do this the way we already work. AI contributes ideas, drafts and decisions. People review the work, correct it and decide what is ready to share. We will document selected examples of what the AI contributed, what people decided and the evidence behind each result. Clinical material will receive expert review.
The first phase is practical and exploratory, over six months to a year. We expect it to produce useful questions and early lessons. It is not a controlled comparison of every approach.
It worked if we publish research others can read and question, and if the AI we build shows choices we can study and explain.
At the end of the funding period, which is six months to a year, we will report what was written, what was built, what the AI contributed, where people stepped in and what it cost. We will share accessible resources and public findings, including the approaches that failed.
The goal is $150,000. The minimum is $50,000. The funding period is six months to a year.
The minimum pays for three things:
Compute, so we can build and test new AI systems.
- Infrastructure to run, store and share the work.
- Bringing new people on, so the research is not limited to what two people can do.
Funding above the minimum would expand the same work: more compute, more people and more research time.
These are planning amounts. Rates, outside specialists and tools will be estimated from real quotes or actual charges before they are spent.
Resonant Labs may also develop commercial products. We will be clear with funders about which work and outputs their support covers. The public research findings proposed here are intended to be free.
Resonant Labs LLC is a women-owned, women-run company. Tab is CEO and leads company operations and program administration.
I'm Brooke Schrickel, a neuroradiologist and educator. I lead the scientific, clinical and educational work.
Paid engineers, clinical reviewers, editors and other specialists will contribute where their expertise is needed.
Our track record is the work already public. We built these projects to put our ideas into practice, and people can explore them now:
- [NeuroRad.ai](https://neurorad.ai/), a free neuroradiology evidence resource. AI helps organize the literature; clinical experts check the evidence.
- [NeuroRad Tracts](https://neuroradtracts.com/), teaching videos for residents and fellows.
- [Learn AI with ChattyG](https://learnaiwithchattyg.com/), AI-created education about using AI, with human oversight.
- [CNSAtlas](https://cnsatlas.ai/), anatomy and imaging teaching.
- AI Office, a working prototype that keeps a person's project context and unfinished work available to compatible assistants.
Resonant Labs also runs the projects below. Each has its own funding request, separate from this one, so this proposal does not pay for their deliverables and does not combine those asks:
- [NeuroRad.ai](https://neurorad.ai/): a free neuroradiology evidence resource. It already has its own [Manifund proposal](https://manifund.org/projects/neuroradai-keeping-neuroradiology-evidence-current-nnmmzqr4b8o).
- [NeuroRad Tracts](https://neuroradtracts.com/): teaching videos for residents and fellows. It already has its own [Manifund proposal](https://manifund.org/projects/neurorad-tracts-).
- [CNSAtlas](https://cnsatlas.ai/): anatomy and imaging teaching. It already has its own [Manifund proposal](https://manifund.org/projects/ai-generation-of-complex-3d-anatomy-and-clinical-understanding).
- Clinical Reasoning: a study of whether expert guidance can improve AI-assisted radiology reporting. Requested separately.
- [Learn AI with ChattyG](https://learnaiwithchattyg.com/): AI-created education about using AI, with human oversight. Requested separately.
This proposal funds the lab's research and new AI only.
The research may take longer than planned. AI changes quickly, and a question we start with may not hold up. If so, we will say what changed and what we learned.
We may find that the AI does not develop taste in a way we can measure. That is still a useful result, and we will publish it.
If we raise only the minimum, we will focus on compute, infrastructure and bringing new people on, and narrow the research scope to match. The report will say what was completed and what remains.
Resonant Labs is self-funded. We have not received grant funding.