LessWrong (30+ Karma) “Publishing academic papers on transformative AI is a nightmare” by Jakub Growiec
Nov 3, 2025
Jakub Growiec is a Professor of Economics who transitioned from growth theory to the complexities of transformative AI. He discusses the struggles of publishing his impactful paper on existential risk, detailing the paradox of widespread conference interest contrasted with repeated journal rejections. Growiec argues for the necessity of subjective probabilities in assessing AI threats, and he critiques the academic tendency to avoid discussions about existential risk due to psychological aversions. He calls for more mainstream research to better inform policy on AI's implications.
AI Snips
Chapters
Transcript
Episode notes
Well-Received Paper, Rejected By Journals
- Jakub Growiec and Klaus Prettner wrote a paper on the economics of p(doom) that was well received at conferences.
- The same paper was desk-rejected seven times by journals despite positive seminar feedback.
Small Probabilities Still Warrant Action
- Referees objected that probability assignments for doom are subjective and lack empirical support.
- Growiec argues the results do not rely on specific probabilities and small extinction risks still justify more investment in risk reduction.
Theory Needs Abstract Concepts
- Reviewers criticized abstract definitions like corrigibility for lacking actionable implementation pathways.
- Growiec notes alignment remains unsolved, so abstract definitions are necessary for theoretical work.