THE “FAKE” PAPER CHALLENGE: CAN AI SPOT BAD SCIENCE?.

Have you heard of “Bixonimania”?

If not, don’t worry… it doesn’t exist!

At our May Journal Club, the committee secretly challenged us to read a ‘scientific paper’ that was completely fake but generated by medical researchers to appear credible. We’re proud to say we spotted it, although it wasn’t quite as obvious as expected.

The purpose? To challenge the way we interact with scientific papers and our use of AI as a tool for digesting information.

THE CONTEXT:

Before each Journal Club, the committee sends out the paper we’ll be discussing so attendees can read it and gather their insights ahead of the session. AI is often used by participants to help interpret the papers we choose whether because we’re a busy agency short on time or because it can make complex research easier to understand.

It’s a convenient tool that combines efficiency with knowledge. However, the committee wanted to see whether people could identify this fake research paper on their own, or whether AI, if used, would spot it too.

Thankfully, everyone who attended and shared their thoughts during the session recognised that something wasn’t quite right.

THE PAPER:

Bixonimania is a fictional preprint(an early draft of a scholarly paper shared publicly before formal peer review) by Lazljiv Izgubljenovic, Betsy Thurberg, and Andi Deep, published on Preprints.orgin May 2024. Written in the style of a randomised controlled trial, it humorously explores a made-up condition allegedly caused by blue-light exposure. See linked here.

Whether it was the AI-generated images of symptoms, the trials taking place at the “Department of Machine Evolution and Human Antics” at the “University of Fellowship of the Ring”, the study involving “fifty made-up individuals”, or the lead researcher, “Professor Sideshow Bob”; attendees were left scratching their heads and wondering…

What is Journal Club up to?!

What made this paper particularly interesting was not only was it entirely fabricated, but that its purpose was to test AI software’s ability to identify and spread misinformation.

Medical researchers invented the skin condition discussed in the paper, ‘Bixonimania’, and uploaded two fake studies to preprint servers. The experiment tested whether large language models (LLMs) would absorb misinformation and repeat it as credible health advice. Within weeks, major AI systems had begun referencing the invented condition. At one point, if you entered symptoms into these LLM platforms, there was a good chance you could be ‘diagnosed’ with Bixonimania.

More troubling still, the condition was even cited in peer-reviewed literature.

Copilot and Gemini described ‘Bixonimania’, with Gemini linking it to blue light exposure. Perplexity AI claimed it affected “1 in 90,000 people”, while ChatGPT assessed whether users’ symptoms matched the condition.

With this in mind, we then moved on to hear from our very own Ami Desai, who shared her top five tips for reviewing scientific papers, from her time in cancer research.

Because, as much as AI can support and streamline how we process information, nothing beats having the real thing in your back pocket.

We all came away with valuable insights and, while we don’t expect to stop using AI, we concluded that combining it with careful factchecking and the foundation provided by Ami’s tips will help us interpret healthcare and medical research findings as accurately as possible.

Who we are

Verve is a health communications agency with the belief that great scientific communication doesn’t just inform – it connects, inspires, and drives action. Science is more than just clinical data and research papers; it’s also about the people it impacts and the ideas that shape the world around us.

Our philosophy at Verve, Science, humanised, brings together the precision of science and the power of human connection. We blend data with empathy, translating complexity into communication that resonates. Whether we’re creating compelling healthcare narratives, simplifying scientific breakthroughs, or engaging audiences with thoughtful storytelling, our mission is the same: to bridge the gap between knowledge and understanding.

Because when science is understood, it becomes more than knowledge, it becomes meaningful. We don’t just convey facts; we create experiences that make science accessible, relevant, and engaging. Our approach ensures that every story is rooted in accuracy but delivered with emotion, inspiring people to take action.

In a world where science is more important than ever, clear and impactful communication is key. Let’s make science something individuals don’t just read about-but truly connect with.

Thanks for reading! Stay curious, stay connected – We give you Verve.

You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy • You might also enjoy

THE “FAKE” PAPER CHALLENGE: CAN AI SPOT BAD SCIENCE?

June 15, 2026

What happens when a fake scientific paper looks convincing enough to fool AI? At May’s Journal Club, we put critical thinking to the test with “Bixonimania”, a completely made-up condition that revealed just how easily misinformation can spread when it appears credible.

HANTAVIRUS EXPLAINED: WHAT WE KNOW SO FAR…

May 8, 2026

Cookies information

We use cookies to understand how visitors use our site. By clicking the “Accept” button, you agree to our use of cookies and similar technologies. You can manage your browser settings to disable cookies at any time. For more information, see our Cookie Policy.

DECLINE