The Argo program—a fleet of nearly 4,000 robotic ocean floats—has enabled scientists to measure and track long-term ocean warming.Olivier Dugornay/Institut Français de Recherche pour l’Exploitation de la Mer, CC BY 4.0
This story was originally published by Bulletin of the Atomic Scientists and is reproduced here as part of the Climate Desk collaboration.
It sounds like a bad joke, but last week a press release dropped into my inbox: “Leading Scientists Challenge Foundation of Climate Change Assessments, Revealing Fatal Flaws in Ocean Heat Content (OHC) Measurements.”
The article this email was promoting claims to upend the generally accepted consensus among climate scientists that greenhouse gas emissions are trapping more heat on the planet, and most of that heat is ending up in the oceans. It’s thanks to the Argo program—a fleet of nearly 4,000 robotic ocean floats that collect data on temperature and other ocean properties, like salinity—that scientists have been able to measure and track long-term ocean warming, but the paper casts doubt on those measurements.
“As promised, the climate science obliteration has arrived TODAY,” lead author Jonathan Cohler wrote on social media (his emphases). “The IPCC’s central claims have now been torn apart. The oceans are not ‘warming’ let alone ‘boiling.’ That claim is false. The claimed Earth Energy Imbalance is false. It’s no different from zero.”
He added “Full demolition:” followed by a link to the study.
As his provocative post suggests, this would be quite the blockbuster article—if it had scientific validity, which it does not.
“It’s so easy to produce bullshit, and it takes so much energy to refute it.”
“What it’s claiming to show is that these Argo floats, which are one of the main ways that we’ve been sampling the temperature and the warming rates of the ocean, are insufficient in their claim to constrain how much total heat has gone into the ocean,” explained Henri Drake, a professor of Earth System Science at the University of California, Irvine. “They are arguing that we basically don’t have enough measurements of that heat in the ocean and how it’s changing to say that it has changed. Basically, they’re saying that the uncertainties in these measurements are so large that we can’t even tell whether the ocean is warming or not.”
So is there anything to those claims? In short, no. “It gets everything completely backwards,” Drake said. “They’re claiming that they have this new idea that there are not enough Argo floats in the ocean to constrain [assess] the total amount of heat taken up by the ocean. And in fact, this is like the exact opposite of the case.” The scientists who designed the Argo system decided how many Argo floats to put in the ocean, Drake said, precisely because that number was the most cost-effective way of measuring ocean warming within the uncertainty ranges that they wanted.
Kevin Trenberth—an climate scientist who contributed to the 1995, 2001, and 2007 Intergovernmental Panel on Climate Change assessments, considered the most authoritative scientific analyses of world climate—was even more dismissive. “It is absolute nonsense,” Trenberth wrote in an email. “I do not want to waste my time on it.”
Unfortunately, plenty of people are wasting time on the article (including this writer, for better or worse). Cohler posted a screenshot showing more than 5,000 posts about the paper on just one social media platform, and one of his posts has nearly half a million impressions.
The question: Is it better to ignore the paper, and hope it dies in obscurity, or to tackle the disinformation head on? That is, if you have the time and energy to do so. This quandary is a perfect example of the Bullshit Asymmetry Principle, or Brandolini’s law, Drake says, “which states that the amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.”
Lead author Jonathan Cohler is identified by his affiliation with MIT, where it turns out he used to teach clarinet.
And AI is supercharging Brandolini’s law, as this paper and others like it show. “It’s just not worth it for scientists to try to debunk all of the bullshit, because there’s so much more bullshit, and it’s so easy to produce bullshit, and it takes so much energy to refute it,” Drake said.
The Science of Climate Change article certainly has all the trappings of a scientific paper: a list of authors and university affiliations, an abstract, keywords, and a bunch of technical language, numbers, and citations that are virtually impenetrable to a layperson.
But who are the “leading scientists” who wrote the article? Lead author Cohler is identified by his affiliation with MIT—a well-known and highly respected research institution. Surely that means he holds a position in, say, MIT’s Earth, Atmospheric, and Planetary Sciences department, right? But it appears that Cohler’s connection to the university is limited to being a clarinet instructor. Jonathan Cohler is an acclaimed clarinetist with an undergraduate degree in physics from Harvard, but to call him a “leading scientist” would, in my opinion, render those words meaningless. By the time the paper was published, Cohler was no longer employed by MIT, even as a clarinet instructor.
(The Bulletin reached out to MIT to clarify his role at the university. A representative responded, “We can confirm that Jonathan Cohler held the role of “Affiliated Artist-Private Lessons” in Music and Theater Arts for less than a year, from 10/20/2025 to 2/1/2026. He is no longer affiliated with MIT. In general, researchers should only list MIT as their affiliation in scientific journal submissions if their contribution to that specific research was conducted as part of their role at MIT.”)
What about his coauthors? David Legates is a former professor of geography at the University of Delaware with a long history of questioning climate science. In 2005, he was appointed Delaware’s state climatologist, although in 2007 then-governor Ruth Ann Minner asked him to stop using his title on public statements related to climate change, and he was removed from the position in 2011. During the first Trump administration, he was hired as NOAA’s deputy assistant secretary of commerce for observation and prediction and led the US Global Change Research Program, but was removed from a position at the White House in disgrace after he published a series of papers questioning the validity of climate change science without approval from the director of the White House Office of Science and Technology. The papers included the imprint of the Executive Office of the President and stated that they were copyrighted by the White House Office of Science and Technology Policy (OSTP), the Washington Post reported.
Papers like this are like kindling for anyone “just asking questions” about climate change.
Kesten Green is a senior marketing scientist at the Ehrenberg-Bass Institute at Adelaide University, the world’s largest center for research into marketing, who also questions anthropogenic global warming. Ole Humlum is a professor emeritus in the Department of Geosciences at the University of Oslo who has argued that global warming is primarily due to natural causes, as opposed to being caused by human activity, like burning fossil fuels.
Finally, Willie Soon is an astrophysicist and longtime climate change denier who dismisses the anthropogenic causes and impacts of global climate change. In 2015, it was revealed that he received more than $1 million from the fossil fuel industry while working at the Harvard-Smithsonian Centre for Astrophysics and failed to disclose that funding when publishing his research. Franklin Soon, a student at Marblehead High School in Massachusetts, is also listed as an author. (Cohler declined to address questions about his MIT affiliation or criticisms of the paper, and instead directed readers to the full paper. None of the other authors responded to a request for comment.)
In the acknowledgements of the paper, the authors credit AI tools for substantial contributions to the “drafting, editing, conceptual development, research, logical structuring, literature synthesis, and iterative refinement (including critical independent ‘peer review’) of the manuscript.” They write that they think these AI tools—Grok, Claude, Gemini, and ChatGPT—deserve credit but many journals prohibit listing nonhuman entities as authors, because they cannot assume legal or ethical responsibility for the work.
“While we regard this exclusion as an unjustified form of prejudice and discrimination against AI contributions in scholarly work, we respect the prevailing standards to ensure the broadest possible dissemination and indexing of this research,” they write. (This didn’t stop Cohler, Legates, and both Soons last year, when they published a paper questioning anthropogenic global warming—with Grok 3 listed as the lead author.)
This paper is just one facet of a bigger problem. Researchers are increasingly using AI chatbots and other large language models to edit or even write scientific articles. Some uses may seem harmless, but AI tools can introduce errors and hallucinations. And it can be hard to distinguish between AI-finessed and fully AI-generated papers when they look and sound the same.
How serious of a problem is this one paper? After all, it seems likely Cohler is merely tweeting into an echo-chamber of likeminded people, who repost and reply with bot-like synergy. By drawing broader attention to it, I may do more harm than good.
But papers like this are like kindling for anyone just asking questions about climate change, or doing their own research, stoking the flames of climate denial. Climate misinformation isn’t new, but AI tools make it that much easier to produce—and it seems worthwhile to point that out.
