The Case of the AI-Generated Giant Rat Penis
An illustration featuring a rat with a cross section of a giant penis set off a firestorm of criticism about the use of generative artificial intelligence based on large language models (LLMs) in scientific publishing. The bizarre illustration was embellished with nonsense labels, including one fortuitously designated as “dck.” The article on rat testes stem cells had undergone peer review and editorial vetting before being published in February by Frontiers in Cell and Developmental Biology.
“Mayday,” blared longtime AI researcher and critic Gary Marcus on X (formerly known as Twitter). Vexed by AI’s ability to abet the “exponential enshittification of science,” he added, “the sudden pollution of science with LLM-generated content, known to yield plausible-sounding but sometimes difficult to detect errors (‘hallucinations‘) is serious, and its impact will be lasting.”
A February 2024 article in Nature asked, “Is ChatGPT making scientists hyper-productive?” Well, maybe, but Imperial College London computer scientist and academic integrity expert Thomas Lancaster cautioned that some researchers laboring under “publish or perish” will surreptitiously use AI tools to churn out low-value research.
A 2023 Nature survey of 1,600 scientists found that almost 30 percent had used generative AI tools to write manuscripts. A majority cited advantages to using AI tools that included faster ways to process data and do computations, and in general saving scientists’ time and money. More than 30 percent thought AI will help generate new hypotheses and make new discoveries. On the other hand, a majority worried that AI tools will lead to greater reliance on pattern recognition without causal understanding, entrench bias in data, make fraud easier, and lead to irreproducible research.
A September 2023 editorial in Nature warned, “The coming deluge of AI-powered information must not be allowed to fuel the flood of untrustworthy science.” The editorial added, “If we lose trust in primary scientific literature, we have lost the basis of humanity’s corpus of common shared knowledge.”
Nevertheless, I suspect AI-generated articles are proliferating. Some can be easily identified through their sloppy and flagrant unacknowledged use of LLMs. A recent article on liver surgery contained the telltale phrase: “I’m very sorry, but I don’t
Article from Reason.com
The Reason Magazine website is a go-to destination for libertarians seeking cogent analysis, investigative reporting, and thought-provoking commentary. Championing the principles of individual freedom, limited government, and free markets, the site offers a diverse range of articles, videos, and podcasts that challenge conventional wisdom and advocate for libertarian solutions. Whether you’re interested in politics, culture, or technology, Reason provides a unique lens that prioritizes liberty and rational discourse. It’s an essential resource for those who value critical thinking and nuanced debate in the pursuit of a freer society.