AI Engineer's Homemade mRNA Vaccine for Dogs Sparks Debate: Tumor Shrinks but Not Yet Cured, Regulatory Hurdles in the Spotlight

According to 1M AI News monitoring, Sydney machine learning engineer Paul Conyngham self-taught mRNA vaccine design using ChatGPT and DeepMind's AlphaFold to create a custom vaccine for rescue dog Rosie, who suffers from mast cell cancer. He collaborated with the University of New South Wales (UNSW) RNA Institute on the vaccine design and completed the injection at the University of Queensland's Gatton Veterinary School. After the injection, one of Rosie's tumors visibly shrank, with treating veterinarian Professor Paola Allavena stating, "The tumor has halved in size." The story has recently been widely circulated on social media as "AI cures dog's cancer."

However, according to the original report released by the University of New South Wales, Rosie's cancer is currently progressing, and a cure is still a long way off. Biomedical engineer Patrick Heizer pointed out on X that manufacturing a single mRNA vaccine is "trivially easy" from a technical perspective. The real challenge, both difficult and expensive, lies in proving the vaccine to be safe and effective in a randomized controlled trial, a step that has not yet been completed.

Conyngham himself also noted that the biggest obstacle throughout the entire process was not vaccine design but ethical approval: he spent three months dedicating two hours each night to writing a 100-page ethics approval document, which he described as "harder than making the vaccine." Biological author Ruxandra Teslo, in her analysis, referenced a similar experience from GitLab co-founder Sid Sijbrandij (who explored experimental therapy at his own expense after osteosarcoma recurrence and has not had a relapse since 2025), suggesting that regulatory bureaucracy in early clinical trials is a key bottleneck hindering the implementation of personalized medicine.