A dog's tennis ball-sized tumor shrank by roughly 75% after receiving a personalized mRNA cancer vaccine that her owner helped design using AI tools, including ChatGPT and AlphaFold.
Paul Conyngham, a Sydney-based tech entrepreneur and electrical/computing engineer, was told his rescue dog Rosie had months to live. He's the cofounder of Core Intelligence Technologies and a former director of the Data Science and AI Association of Australia, so he had the technical background to do something unconventional: he started asking ChatGPT how to fight her cancer.
From ChatGPT Prompt to Lab Bench
ChatGPT pointed Conyngham toward immunotherapy and specifically to UNSW Sydney's Ramaciotti Center for Genomics. He spent $3,000 to have both Rosie's healthy DNA and her tumor DNA sequenced, then used AI tools to analyze the results. AlphaFold, DeepMind's protein structure prediction system (which models how proteins fold into 3D shapes based on their amino acid sequence), helped him identify the specific mutations driving Rosie's cancer and flag potential drug targets.
Pall Thordarson, a nanomedicine researcher and director of UNSW's RNA Institute, took Conyngham's AI-assisted analysis and developed a custom mRNA vaccine in under two months. mRNA vaccines work by teaching the immune system to recognize specific proteins - in this case, the mutated proteins on Rosie's tumor cells.
Rosie received her first injection last December. By mid-March 2026, the tumor had shrunk dramatically.
First Personalized Cancer Vaccine for a Dog
This is reportedly the first time a personalized cancer vaccine has been designed for a dog. The approach mirrors what companies like BioNTech and Moderna are developing for human patients, where each vaccine is tailored to an individual's specific tumor mutations rather than using a one-size-fits-all treatment.
The story illustrates something specific about where AI tools are genuinely useful right now: not replacing experts, but helping motivated non-specialists get to the right experts faster and arrive with better data. Conyngham didn't synthesize the vaccine in his garage. He used ChatGPT to navigate an unfamiliar field, AlphaFold to do protein analysis that would have taken a research team weeks, and then handed everything to actual scientists who built the treatment.
Oncologists have urged caution about generalizing from a single case, and they're right to. One dog's response doesn't validate a treatment protocol. But the pipeline itself - consumer AI for research navigation, specialized AI for molecular analysis, human experts for the actual medicine - is a pattern worth watching beyond veterinary oncology.