Engineers' AI-made mRNA vaccine to save dogs sparks heated debate: tumors have shrunk but are not yet cured, regulatory bottlenecks become a focal point.

According to 1M AI News , Sydney-based machine learning engineer Paul Conyngham used ChatGPT and DeepMind's AlphaFold to learn mRNA vaccine design and created a custom vaccine for Rosie, a rescue dog suffering from mast cell carcinoma. He collaborated with the RNA Institute at the University of New South Wales (UNSW) to design the vaccine, which was administered at the Gatton School of Veterinary Medicine at the University of Queensland. After the injection, one of Rosie's tumors shrank significantly; Professor Paola Allavena, the treating veterinarian, stated that "the tumor shrank by about half." This story has recently been widely circulated on social media as "AI cured a dog's cancer."

However, according to the original report published by the University of New South Wales, Rosie's cancer is still progressing and a cure is still a long way off. Biomedical engineer Patrick Heizer noted on X that manufacturing a single mRNA vaccine is technically "trivially easy," but the real difficulty and expense lies in proving that the vaccine is both safe and effective in randomized controlled trials, a step that has not yet been completed.

Conyngham himself pointed out that the biggest obstacle in the entire process was not vaccine design, but ethical approval: he spent three months, two hours every night, writing a 100-page ethics approval document, which was "more difficult than manufacturing a vaccine." Biology author Ruxandra Teslo, in her analysis, cited a similar experience of GitLab co-founder Sid Sijbrandij (who self-funded experimental therapies after osteosarcoma recurrence and has not relapsed since 2025), arguing that regulatory bureaucracy in early clinical trials is the core bottleneck hindering the implementation of personalized medicine.