Call for Stories on Human and AI Teamwork in Mathematics
Two mathematicians, Deanna Needell and Hayden Schaeffer, are collecting stories from researchers who have worked with AI on recent math results. The call is for personal accounts, not data. So it can map how people are using AI in mathematics, but it can’t yet say how well that collaboration works.
In short: Two mathematicians, Deanna Needell and Hayden Schaeffer, are collecting stories from researchers who have worked with AI on recent math results. The call is for personal accounts, not data. So it can map how people are using AI in mathematics, but it can’t yet say how well that collaboration works.
AI has started to show up in real mathematical research. Nobody has yet written down what that work looks like from the inside.
What happened, in plain words
Deanna Needell and Hayden Schaeffer have put out a public call, shared on Terence Tao’s blog, for stories about humans and AI working together on mathematics. They want to hear from mathematicians who took part in recent AI-assisted results. They also want to hear from those whose earlier research made those results possible. Contributors can reach them by email. The organizers say messages will stay private and nothing will be quoted or used without direct permission.
Key points
- A call for stories: The organizers want to document how human mathematicians and AI actually collaborate in recent work.
- Two kinds of contributors: People who took part in AI-assisted results, and people whose earlier work laid the groundwork for them.
- Confidentiality: Messages stay private, and nothing is used or quoted without explicit permission.
Terms explained
- Preprint — A draft of a scientific paper that has not yet been through formal peer review. Example: A student shares a draft report with classmates before the teacher grades it.
Critical analysis
What is encouraging. Documenting these collaborations while they are still new is useful. Much of what happens in a mathematical discovery (the dead ends, the prompts that failed, who had the key idea) never makes it into a published paper. Collecting accounts now could preserve details that would otherwise be forgotten. Crediting people whose earlier work enabled a result also helps, since such contributions are easy to overlook.
What deserves caution:
- Self-selection. People who answer a call like this are likely to be the ones with positive or notable experiences. Failed attempts, quiet disappointments and cases where AI added little are less likely to be reported, so the picture may lean optimistic.
- Stories are not measurements. Personal accounts show what is possible, not how often it happens or how much AI really contributed. Without a clear way to separate the human contribution from the AI’s, “collaboration” can mean very different things.
- Verification is the hard part. In mathematics, the real test is whether a result is correct and independently checked. Accounts of AI help are most useful when tied to results the community has actually verified.
- Privacy cuts both ways. Confidentiality encourages candor, but if much of the material can’t be quoted or shared, the public may see only a filtered summary of what was learned.
- No findings yet. This is an announcement of a project, so any conclusions about what makes human-AI teamwork effective are still to come.
Bottom line: A sensible first step toward understanding AI’s role in mathematics, but its value will depend on how broadly it collects stories, including the unflattering ones, and how transparently the results are reported.
What we still don't know
How much AI actually contributed to these results, how often such collaborations succeed, and what practices make them work. The source announces the request for stories and reports no findings.
Based on reporting from Terence Tao (UCLA). This is an independent explainer, written in our own words with AI assistance; Terence Tao (UCLA) has not reviewed or endorsed it. Read the original for the full details.