On May 8th, 2024, Nature published an article describing the latest iteration of DeepMind’s groundbreaking protein-folding artificial intelligence software: AlphaFold3. Developed by researchers at Google DeepMind and Isomorphic Labs, AlphaFold3 improved upon its predecessors by modeling protein-ligand interactions at far greater accuracy than previous technologies, among other things. But instead of being met with universal acclaim, the article left a bad taste in the mouths of many within the scientific community.

Unlike AlphaFold2, which was open-source, AlphaFold3 was neither open-weight or open-source upon release. Researchers could access AlphaFold3 through a web server with a hard limit of ten predictions per day. The restriction fueled backlash, as Nature’s editorial policy states, “authors are required to make materials, data, code, and associated protocols promptly available to readers without undue qualifications.” Despite this, Nature published the article without requiring open access to AlphaFold3’s underlying code or data.

Angered researchers wrote an open letter to Nature saying the decision “does not align with the principles of scientific progress” and prevents other researchers from rigorously evaluating claims. In fact, one of the authors of the open letter was a reviewer for the AlphaFold3 Nature manuscript. He allegedly made multiple requests for access to the code during the peer review process, but was denied.

In response, Nature published an editorial justifying their decision, emphasizing that a majority of research is privately funded and that journals ought to engage with this work while ensuring research is reproducible and valid. To many, this argument feels “hand-wavy” at best since the reproducibility of research depends on the availability of potentially proprietary resources.

In the end, Google DeepMind researchers felt it was best to release an open-source version of the model. In November of 2024, Nature released a correction to the original article, stating that “AlphaFold3 is now more open.” However, access is still restricted to those with an academic affiliation.

As artificial intelligence-driven research continues to advance, much of the progress is being driven by large tech companies with deep pockets. The tension remains: should the private sector be allowed to benefit from the exposure and clout gained by publishing their work in prestigious journals without fully adhering to the norms of open science? Can the scientific community demand all proprietary work be unveiled for the purposes of transparency without jeopardizing the very funding that fuels such breakthroughs? In the end, neither side can have it both ways–and as the gears of capitalism continue to turn, the question of who truly controls scientific progress remains unresolved.