Desperate to save his sick dog, an Australian man used artificial intelligence to design a personalized experimental treatment, enlisting the help of renowned scientists to administer it.
Eight-year-old Rosie’s mast cell cancer is now in partial remission and her largest tumor has shrunk considerably, her owner Paul Conyngham, an AI consultant in Sydney, told AFP.
“She has regained a lot in mobility and activities” after receiving a personalized mRNA vaccine combined with powerful immunotherapy in December, he said.
This long fight against Rosie’s cancer attracted the attention of OpenAI boss Sam Altman, who mentioned this “incredible story” Friday March 27 on X.
Paul Conyngham does not claim to have found a miracle cure, but this fierce battle highlights, according to researchers, the potential of AI to accelerate medical research.
“I was constantly chatting with ChatGPT, Gemini and Grok to study cancer therapies in depth”a expliqué M. Conyngham.
Following the advice of the chatbots, he paid 3,000 dollars (2,600 euros) to have Rosie’s genome sequenced.
He uses the same online tools to analyze his DNA data, then turns to AlphaFold, a scientific AI model that won the Nobel Prize in Chemistry in 2024, to better understand one of his dog’s mutated genes.
Also thanks to a recommendation from ChatGPT, the Australian is seeking help from a team from the University of New South Wales (UNSW) and other Australian researchers to carry out his work.
Wrong diagnosis
For almost a year, Rosie’s cancer was misdiagnosed, Conyngham said. “I took her to the vet three times. On two occasions he told me not to worry, that it was just a rash.”he explained.
But Rosie’s condition worsens and a biopsy reveals in 2024 that the dog has terminal cancer.
After trying chemotherapy, conventional immunotherapy and surgery, the costs piled up and Mr Conyngham looked for other options.
It uses artificial intelligence to look at emerging treatments, including mRNA vaccines, which boost the immune system and have been widely used during the Covid-19 pandemic.
“The combination of three disruptive technologies – genome sequencing, artificial intelligence and RNA therapies – offers new perspectives and raises new challenges”considers Martin Smith, professor at UNSW, who sequenced Rosie’s genome.
More “this was in no way a clinical trial” et “AI will not cure cancer”, he specifies.
The start of a new era in medical research?
Chatbots helped Mr Conyngham navigate the mountain of paperwork needed and, using his new scientific network, he met with a professor from the University of Queensland to administer the treatment.
But not all tumors responded as well as the largest. Rosie had to undergo another operation and it is unknown how long she has left to live.
“We don’t know exactly.” which made it possible to reduce the size of Rosie’s largest tumor, recognizes Pall Thordarson, director of the UNSW RNA Institute, creator of the vaccine.
Paul Conyngham “used the AI program to design the mRNA sequence. And he then gave us that information“, he explains.
Lack of publication of scientific details by UNSW and Mr. Conyngham, “We don’t know enough about the vaccine to understand how much AI contributed to its development, or even whether the vaccine worked as intended.”explains to AFP Nick Semenkovich of the Medical College of Wisconsin, unrelated to the Rosie affair.
But he recognizes that “AI is very promising” for medical research.
We go from “looking for a needle in a haystack, to a data-driven selection process, which significantly shortens the time between diagnosis and vaccine development” and can help animals and humans to survive, adds Patrick Ming-kuen Tang, professor at the Chinese University of Hong Kong, even if the risk of errors is real.
