OpenAI Researcher Allegedly Pressured Mathematician to Drop Anthropic Co-Author from Math Breakthrough Paper
A prominent mathematician claims an OpenAI researcher pressured him to remove a co-author who works at rival AI company Anthropic from a major mathematics breakthrough paper.
Who: University of Chicago mathematician and Fields Medal winner Prof. Laszlo Baba.
What: Baba alleges OpenAI researcher and AI safety expert Paul Christiano demanded he drop Anthropic employee Yash Sharma as a co-author.
Why: The pressure allegedly stemmed from Sharma’s employment at Anthropic, OpenAI’s direct competitor in developing large language models.
When: The incident occurred during the revision process of a paper detailing a new mathematical framework for understanding neural network generalization.
The paper, which Baba describes as a “once-in-a-decade” insight, was submitted to a top-tier journal. Christiano contacted Baba directly after the initial submission, according to screenshots reviewed by The Decoder.
“He said having Yash on the paper created a ‘reputational risk’ for OpenAI,” Baba told the publication. “He didn’t say it in a threatening way, but the implication was clear: include Sharma, and OpenAI’s support for this research line would end.”
The Alleged Conversation
Baba claims the conversation included specific language about “competitive landscape sensitivity.” Christiano reportedly framed the request as a courtesy to avoid “unnecessary signaling” to the AI research community that OpenAI and Anthropic collaborate.
Baba rejected the request. He stated that Sharma contributed equally to the mathematical proofs and that removing him would constitute scientific misconduct.
“The work speaks for itself,” Baba said. “We are all researchers first. Company affiliations should not dictate authorship.”
Anthropic’s Response
Anthropic declined to comment on the specific allegation. A company spokesperson said, “We support collaborative research across the field. Co-authorship should reflect contribution, not corporate borders.”
OpenAI did not respond to multiple requests for comment. Christiano, who left OpenAI in 2021 but remains a contractor, has not publicly addressed the allegations.
Broader Implications for AI Research
The incident highlights growing tensions between AI labs over research credit and information sharing. Both OpenAI and Anthropic have recently implemented stricter publication policies.
Key Insight: This is the first publicly reported case of a researcher allegedly being pressured to exclude a competitor’s employee from academic authorship. It raises questions about how corporate rivalries are affecting peer review and scientific integrity in AI.
Industry experts warn that such pressure, if substantiated, could chill cross-company collaboration at a time when fundamental breakthroughs require diverse expertise.
Background: The Math Breakthrough
The paper in question tackles the “generalization puzzle” — why neural networks perform well on data they have never seen. Sharma co-developed the core theoretical framework while at Anthropic.
Baba says he will proceed with the original author list. “Science does not recognize corporate boundaries,” he said.
The journal has not commented on the dispute.
The Bottom Line
The alleged pressure to drop an Anthropic co-author from a mathematics paper, if true, represents a significant escalation in corporate influence over academic research. It threatens the collaborative norms that have driven AI’s recent advances. The final outcome will set a precedent for how AI companies treat co-authorship from rivals.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.