Paul Sutter, a cosmologist, is deeply invested in the integrity of science. He studies the origins, structure, and evolution of the Universe. During the COVID-19 pandemic, he watched with alarm as public trust in science diminished.
Academic journals see "staggering profit margins approaching 40%…. Meanwhile, academics do almost all of the substantive work to produce these articles free of charge". Then, we have to pay eye-watering sums to make our papers open access.https://t.co/viYMDV9UR7
— Earl K. Miller (@MillerLabMIT) July 16, 2024
This erosion of trust compelled him to question how to make science more trustworthy. In his latest book, Sutter explores several sources of mistrust.
Excellent piece by Arash Abizadeh explaining the reasons behind PAPA’s entire editorial board resignation (JPP did the same a few months ago) and the exploitation on which (commercial) academic publishing is based. https://t.co/lXQ6bGDZE5
— Lea Ypi (@lea_ypi) July 16, 2024
These include the stigma against scientists who engage with the public, the lack of stable scientific careers, and the complicity of scientists when their work is politicized.
He proposes proactive steps to address these issues and rebuild trust in science. One critical issue is the relentless pressure on scientists to publish.
Everything described quite accurately, we should just stop supporting (=working for free for) extortionist publishing houses. Let's ask for a fair amount for editorial or review work. https://t.co/SWDONX04B1
— Jan Trka (@JanTrka) July 17, 2024
This can encourage various forms of fraud, ranging from outright data fabrication to plagiarism, data manipulation, and selective methods to achieve desired results.
In medical publishing, decades of outrage have brought little change to a perverse but mutually beneficial system
Let’s see if the humanities can be transformative“Academic journals are a lucrative scam – and we’re determined to change that” https://t.co/seOAm4iUBg
— Jocalyn Clark (@jocalynclark) July 17, 2024
Fraud diminishes public trust in science, making it crucial to shift the incentive and reward structures within the scientific community. Modern science is incredibly complex and heavily reliant on computational tools. Virtually every scientific paper, regardless of the field, involves some level of computer work.
Fixing trust in scientific publishing
This reliance makes peer review, a key defense against fraud, increasingly ineffective. Many software codes used in scientific research are not publicly available, depriving reviewers of the ability to scrutinize the tools that generate results.
This lack of transparency in computational science is troubling. Scientists are not incentivized to make their code available because it doesn’t contribute to their academic metrics, such as the h-index. This oversight compromises the peer review process, making it difficult to detect errors or fraudulent activities.
If mistakes are made in a manuscript, reviewers can catch and correct them, improving the quality of science. However, if the errors are in the code, they often go unnoticed, which degrades scientific reliability. As science grows more intricate and dependent on software, the potential for fraud—intentional or accidental—increases.
Unintentional errors can slip through due to the software’s complexity, while intentional fraud can be easily hidden. Both scenarios undermine the integrity of scientific research. Fixing these systemic issues requires a fundamental change in how scientific contributions are valued and incentivized.
It is a daunting task but is essential for the future of science and its role in society.
Cameron is a highly regarded contributor in the rapidly evolving fields of artificial intelligence (AI) and machine learning. His articles delve into the theoretical underpinnings of AI, the practical applications of machine learning across industries, ethical considerations of autonomous systems, and the societal impacts of these disruptive technologies.
