An Australian Staffordshire crossbred named Rosie received an AI-designed mRNA vaccine that dramatically shrank her mast cell cancer tumors, offering a glimpse of personalized medicine's potential while exposing regulatory barriers that could delay human treatments by five to ten years.

How Did AI Help Design a Cancer Vaccine for a Dog?

Rosie was diagnosed with mast cell cancer, a common skin cancer in dogs, about two years ago. After multiple surgeries, immunotherapy, and targeted anti-cancer drugs failed to fully control her condition, her owner Paul Conyngham, a tech entrepreneur, turned to artificial intelligence chatbots to explore alternative treatment options. Using his technical background, Conyngham processed gigabytes of genetic data from both Rosie's healthy cells and her cancer cells through AI models.

The AI models streamlined the design process for an mRNA vaccine, which Professor Pall Thordarson and a team from the University of New South Wales (UNSW) then synthesized and administered to Rosie. The vaccine worked: the tumors on her legs shrank dramatically. While Rosie still has cancer and only has several months to live, the treatment extended her lifespan and improved her quality of life, allowing her to continue running and playing.

"One of the things that we were able to get was the genomic healthy cells from the dog and then from the cancer, and then work out what might be possible to do with that," explained Professor Thordarson.

Professor Pall Thordarson, UNSW

The AI did not cure Rosie's cancer or create the vaccine itself, but it played a critical role by accelerating the vaccine design pipeline from years to months and reducing manufacturing errors.

Why Are Regulatory Frameworks Blocking Human Trials?

Rosie's case has captured global attention, even swaying U.S. Health Secretary Robert F. Kennedy Jr., who declared at a Senate hearing in April 2026 that personalized AI-driven medicine will transform healthcare. However, scientists warn that Australia and other countries face a fundamental mismatch between the speed of AI-driven drug design and the pace of regulatory approval.

In human medicine, each personalized cancer vaccine requires DNA sequencing unique to the patient's tumor. The manufacturing process for each vaccine is nearly identical, yet every single one must undergo the same regulatory approval process from the beginning, which can take between five and ten years.

"It's sort of heartbreaking, because I get quite a few messages from people asking whether we can do anything for their loved ones and people ask the same thing about their animals. We're just not in a position as this was a recent project and you know, to go to that scale requires time. That's when we have to start to talk about these regulatory environments," said Professor Thordarson.

Professor Pall Thordarson, UNSW

Professor Thordarson emphasized that regulatory and reimbursement frameworks were established "50 to 100 years ago" and were never designed for personalized, AI-accelerated treatments. He argued that updating these systems is not just a scientific challenge but a political one.

What Three Technologies Converged to Make This Possible?

• RNA Technology: mRNA vaccines can be rapidly designed and synthesized to target specific tumor mutations, making them ideal for personalized cancer treatment.
• Genomics: DNA sequencing of both healthy and cancer cells provides the biological blueprint that AI models use to identify tumor-specific targets.
• Computation: AI models process massive genomic datasets to design vaccine sequences and optimize manufacturing, compressing years of work into months.

Together, these three fields created a pathway for personalized medicine that was previously impossible at scale.

What Do Experts Say Australia Must Do to Lead in Precision Medicine?

Martin Smith, from UNSW's Ramaciotti Centre for Genomics, emphasized that Rosie's case demonstrates more than just a successful animal treatment. It shows that AI can democratize cancer biology, allowing a non-expert tech entrepreneur to understand complex genomics and generate the code needed to process tumor data.

"What's really compelling about the story is it's not just that it's a dog, or the guy with no technical, real expertise in cancer or biology. It's the fact that we can shorten what used to take years and millions of dollars into a more realistic and practical treatment of cancer," said Martin Smith.

Martin Smith, Ramaciotti Centre for Genomics, UNSW

Smith called on the medical community to support this research and make the technology more accessible. He noted that Australia is currently lagging in cancer genomics and that while the government has invested in infrastructure and grants, the real gap lies in operationalizing these clinical projects into practice.

Professor Thordarson argued that Australia now has a unique opportunity to take the lead in precision medicine by modernizing its regulatory frameworks to accommodate AI-designed, personalized treatments. "It's almost a political decision at some point. This is where Australia could take the lead," he stated.

Professor Thordarson

How Can the Medical Community Support AI-Driven Precision Medicine?

• Regulatory Reform: Update drug approval processes to recognize that identical manufacturing procedures for personalized vaccines should not require separate five to ten-year approval cycles for each patient.
• Infrastructure Investment: Expand access to DNA sequencing, computational resources, and mRNA manufacturing facilities so that personalized vaccines can be produced at scale and cost-effectively.
• Research Funding: Move beyond small grants and infrastructure support to fund the operationalization of clinical projects, bridging the gap between proof-of-concept and widespread patient access.
• Cross-Sector Collaboration: Encourage partnerships between academic researchers, biotech companies, and government agencies to standardize protocols and reduce redundant regulatory burdens.

Rosie's story is not a clinical trial but rather a "citizen science project that showcases what's possible," according to Smith. The challenge now is translating that possibility into reality for human patients facing cancer.