Stanford professor challenges DeepMind: protein watermark paper falls short of biosecurity claims
Stanford professor Anshul Kundaje and DeepMind's Pushmeet Kohli engaged in a public dispute over a protein sequence watermarking paper, with Jeremy Ratcliff also joining the debate. Kundaje's core position: the algorithm is interesting and the work may have future biosecurity value as a pilot study, but the paper's evidence does not support the biosecurity claims being promoted.
Confirmed
- Kundaje affirmed the value of the algorithm and paper itself; his critique targets the gap between conclusions and promotion, not the research direction.
- He noted that while the paper's discussion section acknowledges limitations, evidence from Rohit et al. shows that, given the results presented, removing the watermark is trivial—raising the question of what practical biosecurity use it serves.
- Kundaje urged the team to release a preprint like other research, stressing that scientific review beyond the three journal reviewers matters and should not be dismissed.
- The debate also touched on academic evaluation norms: Kundaje argued that in the AI era, expecting young scientists to rely solely on journal review is outdated; he and Jeremy Ratcliff clashed over impact framing and preprint timing.
Why it matters
- The dispute highlights a general problem in AI for Science: the gap between promotional claims and actual evidence strength, especially acute under high-stakes narratives like biosecurity.
- If watermarks can be easily erased, their value as a biosecurity defense is questionable, directly affecting how this research direction is positioned and funded.
- Kundaje's call for preprints and open review continues the ongoing trend of rapid open peer review replacing traditional journal review in AI research.
2026-10-04 ~ 2026-10-04 · 6 related posts
- Episode 1: DeepMind Unveils SynthID Bio, Open-Source Watermarking for AI-Designed Proteins(2026-09-30, 14 posts)
- Episode 2: Google's SynthID Bio Protein Watermarking May Generalize Beyond AlphaFold 3(2026-10-01, 3 posts)
- Episode 3: Researchers Debate Whether Protein Watermarks Can Survive Redesign(2026-10-04, 3 posts)
- Episode 4: Stanford professor challenges DeepMind: protein watermark paper falls short of biosecurity claims(2026-10-04, 6 posts)
- Episode 5: DeepMind Unveils SynthID Bio Watermarks for AI-Designed Proteins(2026-10-04, 4 posts)
Primary sources
- Stanford's Kundaje: Telling junior scientists to rely only on journal review is backward in the AI age — anshulkundaje ·
- Stanford professor challenges DeepMind biosecurity watermark paper as trivially removable — anshulkundaje ·
- Kundaje urges DeepMind to post preprints, defends open scientific critique — anshulkundaje ·
- Stanford Prof. Anshul Kundaje doubles down: DeepMind's paper results don't back biosecurity claims — anshulkundaje · 2026-10-04
- [source] Stanford's Kundaje: Telling junior scientists to rely only on journal review is backward in the AI age — anshulkundaje · 2026-10-04
- [source] Stanford professor challenges DeepMind biosecurity watermark paper as trivially removable — anshulkundaje · 2026-10-04
- Researcher questions erasable protein watermarks as a biosecurity measure — anshulkundaje · 2026-10-04
- Stanford's Kundaje disputes biosecurity framing of a study, argues preprints beat frozen publications — anshulkundaje · 2026-10-04
1 near-duplicate retellings: anshulkundaje