On September 28, 2026, French investigative newspaper Mediapart reported antisemitic comments by far-right leader Jordan Bardella. Reporting on the issue, Le Monde quoted the vice president of the house Sébastien Chenu saying of Bardella's comments, "tout ça est faux" ("this is all made up.")

I call bullshit.

Bullshit (or BS) has been used in philosophy since the 1980s when Harry Frankfurt defined it as (paraphrasing) information that is misleadingly impressive, important, informative, or otherwise engaging.

Now, an ability to detect BS has always been useful, but I'll contend that ours is the era of turbocharged BS, because of generative AI / LLMs. See, an LLM is the ultimate BS generator, creating a compelling argument irrespective of whether it has evidence to support it or not. I'm not saying that LLMs are completely useless, I'm saying that, like a sleazy used-car dealer, they're really good at sounding fantastic irrespective of whether they're selling you the latest Mercedes Benz or a beat-up 1970s Citroën Méhari (it's like a 2CV but made of cheap plastic) way past its due time.

The era of turbocharged BS

You might have seen a meme going around a few months ago, "I need to get my car washed but the car wash is only 300 meters away. Should I walk or drive to the car wash?" To help me answer this critical question, I put my LLM in research mode—you know, the one that burns many more tokens (and trees) and takes longer but that brings in additional quality—and let it work out its magic. After a few minutes, it shared its conclusions: Walk to the car wash, drive your clean car back.

LLMs are the ultimate BS generators in that they produce convincing arguments irrespective of the quality of the evidence.

Not only that, but my LLM had wonderful arguments for supporting its non-sensical conclusions: I'd get some fresh air, save gas money, spare the environment, and avoid parking hassles! It even had further considerations, telling me about the best approach and when to deviate from the recommendation.

Blink and you'll miss it. The argumentation is so strong that if you use, say, an inappropriate border of proof (asking yourself "can I believe this?" instead of "must I believe this?") you'll end up accepting LLM outputs that you should not.

All told, that was very impressive. Except that it was completely nonsensical, not unlike Nobel laureate Richard Feynman's Cargo Cult science,

"In the South Seas there is a Cargo Cult of people. During the war they saw airplanes land with lots of good materials, and they want the same thing to happen now. So they’ve arranged to make things like runways, to put fires along the sides of the runways, to make a wooden hut for a man to sit in, with two wooden pieces on his head like headphones and bars of bamboo sticking out like antennas—he’s the controller—and they wait for the airplanes to land. They’re doing everything right. The form is perfect. It looks exactly the way it looked before. But it doesn’t work. No airplanes land. So I call these things Cargo Cult Science, because they follow all the apparent precepts and forms of scientific investigation, but they’re missing something essential, because the planes don’t land."

So, with the advent of LLMs we're now under a stream of BS, constantly having to ask ourselves not just, "can I believe this" but "must I believe this." The difference is nontrivial, because the latter might help you catch some confirmation biases that the former won't. Defaulting to a more critical processing of comments displays the healthy skepticism that underpins the scientific method, and we could do a lot worse than using it in an age where we need to constantly question what we see.

The lady doth protest too much, methinks

So why is Chenu's "this is all made up" comment BS? Well, from an epistemology standpoint, the only way for Chenu to know that Bardella never made these comments would be to know all the comments that Bardella ever proffered or wrote. That's just not happening.

Not only that, but the strong rejection is in itself a signal. See, Chenu could have gone plenty of different ways, from "this strikes me as wrong" to "if Bardella said so, he was very young at the time and, surely, he is now of a different opinion." But no. He chose to make a blanket rejection that he cannot uphold from a logic standpoint and that should trigger further skepticism on our side.

Ours is the era of turbocharged BS, and the only way we'll thrive in it is by engaging our critical thinking—asking better questions ("must I believe this?")—and using the training that people who study rhetoric use—why is this person telling me this? why now? why in those terms? This is an era where we can all learn from philosophers and humanists who deeply trained to decode arguments. And, for the rests of us—like this engineer—we'd better pick up the skills on the double, 'cause that train has left the station and it's not waiting for anyone.

Re-engaging our critical thinking is especially true when it applies to arguments that we like, that we want to hear. Quoting Feynman's Caltech 1974 commencement address again, "The first principle is that you must not fool yourself—and you are the easiest person to fool."

As for Bardella and his far-right acolytes, they can kindly go f*ck themselves.

References

Berteau, Alexandre, and Antton  Rouget. 2026. '« Les banques sont toutes détenues par des juifs » : Jordan Bardella face à ses écrits antisémites', Mediapart, 28 september 2026.

Feynman, Richard P. 1998. 'Cargo Cult Science', Engineering and Science, 37: 10–13.

Frankfurt, Harry G. 2005. On bullshit (Princeton University Press).

Hicks, Michael Townsen, James Humphries, and Joe Slater. 2024. 'ChatGPT is bullshit', Ethics and Information Technology, 26: 1–10.

Iacobucci, Serena, and Roberta De Cicco. 2022. 'A literature review of bullshit receptivity: Perspectives for an informed policy making against misinformation', Journal of Behavioral Economics for Policy, 6: 23–40.

Flyvbjerg, Bent. 2025. 'AI as Artificial Ignorance', Project Leadership and Society.

Littrell, Shane. 2026. 'The Corporate Bullshit Receptivity Scale: Development, validation, and associations with workplace outcomes', Personality and Individual Differences, 255: 113699.

Tigard, Daniel W. 2025. 'On bullshit, large language models, and the need to curb your enthusiasm', AI and Ethics: 1–11.