ChatGPT is not a solution for canine cancer treatment.
When an Australian tech entrepreneur with no medical or biological background claimed that ChatGPT helped save his dog from cancer, the story quickly gained traction, showcasing the potential of AI in revolutionizing medicine and tackling one of its most formidable challenges. However, the reality is more nuanced.
The narrative that circulated online, initially reported by an outlet, was fairly straightforward. In 2024, Paul Conyngham from Sydney discovered that his dog, Rosie, had cancer. While chemotherapy managed to slow the progression of the disease, it didn’t reduce the tumors. After veterinarians indicated that “nothing could be done” for his Staffordshire bull terrier-shar pei mix, Conyngham took it upon himself to seek a solution.
He turned to ChatGPT for brainstorming treatment options, which suggested immunotherapy and directed him to experts at the University of New South Wales. They conducted a genetic profile of Rosie’s cancer, and Conyngham utilized ChatGPT along with Google’s AlphaFold, an AI model for protein structure, to interpret the findings. Collaborating with UNSW professor Pall Thordarson, he developed a personalized mRNA vaccine tailored to Rosie’s tumor mutations. Thordarson noted that this might be the first instance of such a treatment being designed for a dog.
A few weeks after Rosie received her first injection last December, Conyngham reported that her tumors had shrunk, and she was more active, even chasing rabbits in the park. However, not all tumors responded, and one remained unchanged. “I’m under no illusion that this is a cure, but I do believe this treatment has bought Rosie significantly more time and quality of life,” he stated.
As the story spread, the nuances were often overlooked. Some headlines proclaimed that a non-medical owner had “invented” a cure for his dog’s terminal cancer, while others celebrated the case as a breakthrough in personalized medicine. Prominent figures in tech, including the president of OpenAI, should have known better, while others, like the CEO of Google DeepMind, shared the story without embellishment. Elon Musk also chimed in, highlighting the role of xAI’s Grok, a detail that was often omitted in the initial coverage.
The narrative gives AI too much credit. Rosie was not cured of cancer, and it remains uncertain whether the mRNA vaccine contributed to her improvement. The personalized treatment was administered alongside another immunotherapy known as a checkpoint inhibitor, complicating the ability to determine the vaccine’s effectiveness. One of the scientists involved, Martin Smith, mentioned that tests are being conducted to assess the immune response.
Moreover, the vaccine itself wasn’t generated by a chatbot. ChatGPT did not design or create Rosie’s treatment; it acted more as a research assistant, helping Conyngham navigate medical literature—an impressive feat, but far from the groundbreaking innovation suggested.
Details regarding AlphaFold’s involvement are also vague. A professor from the University of Queensland explained that while the model could provide structural hypotheses about proteins, it is not a comprehensive cancer-vaccine design system. Official guidance indicates that AlphaFold is not validated for predicting the effects of certain mutations and does not account for several biologically significant contexts.
The contribution of Grok is even less clear. Conyngham mentioned on social media that “the final vaccine construct for Rosie was designed by Grok,” but the specifics of that process remain ambiguous. Experts suggest that Grok, like ChatGPT, is a tool that can assist with literature searches, summarizing research, translating jargon, and suggesting workflows—helpful, but not equivalent to designing a cancer vaccine.
Overall, experts believe Rosie’s case should be viewed as an unusual, highly specific proof of possibility rather than a replicable template for the average person. It required substantial expert input, not just a chatbot and a few prompts.
This distinction is crucial in medicine, where success hinges on not only generating plausible information but also on the expert work involved in producing, testing, and delivering actual treatments. An assistant professor in Singapore emphasized that framing this as an AI-driven success overlooks the significant human effort required, without which the AI’s output would remain mere text.
The entire situation carries a hint of a PR stunt, with bold claims built on shaky foundations and vague methodologies that fit neatly into the tech fundraising narrative. mRNA vaccines, much like the broader promise of personalized medicine, are still largely unproven as cancer treatments in humans, let alone in dogs. While the case may be genuine, it glosses over the substantial financial and expert resources necessary to transform an idea into a viable treatment.
I reached out to Conyngham for a discussion but have yet to receive a response. His profile states “Ending Cancer for Dogs” and links to a form outlining his “dream to make this process accessible to everyone.” The form inquires whether your dog has cancer, if you’re a researcher or scientist interested in participating, and if you are an investor.
While it would be a mistake to dismiss Rosie’s story entirely, it’s important to recognize that AI isn’t set to replace traditional lab work anytime soon. However, it is making scientific knowledge more accessible to everyday people. That said, accessibility to care remains a significant issue, as few patients or pet owners have the necessary access to top-tier experts, specialized equipment, and the funds required to turn that information into real treatments.
