Thom discloses a related episode, questions OpenAI researchers' academic standards
Andreas ThomMathematician, TU Dresden
Mathematician Andreas Thom, in a guest post on Terence Tao's blog, said he asked OpenAI researchers Mark Sellke and Sébastien Bubeck whether conversations he and a colleague had with ChatGPT about his research were part of the training data or accessible to the model's reasoning process, and that Sellke denied it. Thom calls that denial materially misleading in light of OpenAI's later statement in the Buckmaster-Alpöge case that it could not rule out that de-identified usage data helped improve its models. He also writes that Bubeck's acknowledged career-related remark to Buckmaster and his objection to including Alpöge in a proposed paper presenting OpenAI's proof "deepen my concern about their commitment to academic standards."
“deepen my concern about their commitment to academic standards”
Original text
Where this statement fits
Did OpenAI improperly pressure mathematicians over authorship and credit?
Tristan Buckmaster describes pressure over authorship and publication during discussions about competing AI-assisted mathematics results. Sébastien Bubeck disputes that characterization and says the proposed authorship concerned OpenAI's own proof. His apology for career-related wording does not settle the disagreement. Researchers and commentators divide over whether the negotiations were appropriate, while raising a separate dispute about the treatment of unpublished user research.
Source and context
Original text
About this source
Guest post by Andreas Thom on Terence Tao's blog, disclosing a related episode involving OpenAI researchers Mark Sellke and Sébastien Bubeck and unpublished conversations about a separate math problem.
Before the quotation
Guest post on Terence Tao's blog disclosing a related episode after OpenAI's August "non-sofic groups" result.
After the quotation
Thom's account concerns a different problem (non-sofic groups) than Navier-Stokes, but bears on the same researchers' conduct toward unpublished outside research.
How this statement is classified
The label describes this statement's response within the context above.
Why this label?
Thom calls Sellke's denial materially misleading and says Bubeck's acknowledged career-related remark to Buckmaster and his objection to including Alpöge in a proposed paper "deepen my concern about their commitment to academic standards," adding that these episodes together raise serious questions about their judgment and personal integrity. That is a direct condemnation relevant to the case's pattern-of-conduct question.
- Recorded on
- Published here
- People and groups discussed
- Sébastien Bubeck, OpenAI
Corrections to this statement
These notices document reporting corrections, classification changes, and topic-label updates.
Reporting correction
Correction to: Statement by Andreas Thom
- Previous language
- Sellke denied it -- a denial Thom argues was later shown misleading by OpenAI's own admission in the Buckmaster-Alpöge case -- saying this "deepens my concern about their commitment to academic standards."
- Corrected language
- Thom calls that denial materially misleading in light of OpenAI's later statement in the Buckmaster-Alpöge case that it could not rule out that de-identified usage data helped improve its models. He also writes that Bubeck's acknowledged career-related remark to Buckmaster and his objection to including Alpöge in a proposed paper presenting OpenAI's proof "deepen my concern about their commitment to academic standards."
An earlier version of this entry said Sellke's denial led Thom to write that it "deepens" his concern about academic standards. In his post, Thom ties that concern to Bubeck's career-related remark to Buckmaster and to his objection to including Alpöge in the proposed paper. We corrected the summary and restored Thom's exact words.
This corrects our account of what happened.
Guest post by Andreas Thom on Terence Tao's blog, disclosing a related episode involving OpenAI researchers Mark Sellke and Sébastien Bubeck and unpublished conversations about a separate math problem.
More from this case
Read the full caseAGMAI calls OpenAI's release a first step, not an endorsement
Advisory Group on Mathematics and Artificial Intelligence
“AGMAI’s advisory role should not be interpreted as a judgment of the impact of these results or an endorsement of the process by which OpenAI obtained them.”Read statement
OpenAI describes steps to improve mathematical disclosures
OpenAI
“We will continue to act on feedback from the community and update our standards for future disclosures of major scientific advancements.”Read statement
AGMAI opposes using mathematical releases as marketing
Advisory Group on Mathematics and Artificial Intelligence
“We strongly recommend that AI labs refrain from treating the release of mathematical results as marketing vehicles to promote their models”Read statement
Share or embed this statement
The embed keeps its attribution link to The Dispute Index. Credit the image to The Dispute Index and link to the statement page.
Cite this record
- Publisher
- The Dispute Index
- Title
- Thom discloses a related episode, questions OpenAI researchers' academic standards
- First published
- Last updated
- Permalink
- https://disputeindex.com/events/829-mathematician-andreas-thom-in-a-guest-post-on
Last updated marks the most recent saved version of this published statement.
The Dispute Index. "Thom discloses a related episode, questions OpenAI researchers' academic standards". First published: 2026-09-17. Last updated: 2026-09-25. https://disputeindex.com/events/829-mathematician-andreas-thom-in-a-guest-post-on