A feel-good story has been circulating about Paul Conyngham, a machine learning professional who reportedly used AI to develop a personalized mRNA cancer vaccine for his dog Rosie. The narrative - "man uses ChatGPT to cure his dog's cancer" - is a perfect viral headline. It's also deeply misleading about what AI actually contributed.
Here's what actually happened. Conyngham, who has 17 years of experience in machine learning and data analysis, worked with researchers at the University of New South Wales (UNSW) to design a personalized mRNA vaccine targeting his dog's mast cell cancer. The process followed five well-established steps: sequencing the tumor's DNA and RNA, identifying candidate antigens (proteins the immune system can target), selecting the best candidates based on immunogenicity (how strongly they trigger an immune response), designing mRNA sequences, and manufacturing the vaccine. The treatment worked. Rosie's cancer responded.
What ChatGPT Actually Did
Conyngham used ChatGPT to explain oncology concepts and assist with data analysis tasks. That's meaningful - having a tool that can translate complex medical terminology into plain language and help process data is genuinely useful. But ChatGPT did not design the vaccine. It did not replace the UNSW researchers who handled the large bulk of the actual scientific process. It did not eliminate the need for Conyngham's own 17 years of technical expertise.
Strip away the viral framing and you get a story about a technically skilled person with institutional access to a major research university using established mRNA vaccine techniques - the same approach companies like Moderna use - with ChatGPT as one tool among many.
The Real Lesson for AI Users
This story matters because it represents how most practical AI use actually works versus how it gets reported. AI as a research assistant, explainer, and data processing aid is genuinely powerful. AI as a replacement for domain expertise, institutional resources, and years of scientific training is fiction.
The average person cannot replicate this outcome. Not because ChatGPT isn't accessible, but because the other 95% of the process requires specialized knowledge, lab access, and collaboration with trained researchers. Framing it as "AI cured a dog" strips out every part that actually mattered.
For those of us who use AI tools daily, this distinction is critical. ChatGPT and similar tools are at their best when they augment real expertise - helping you work faster, understand new domains quicker, and process information more efficiently. The moment you expect them to replace expertise entirely, you're buying into the same hype that turned a collaborative scientific effort into a misleading headline.