Coolest meeting this week: Sam Altman meets man behind ChatGPT-built dog cancer vaccine

It seems AI is now at the centre of everything, thanks in large part to ChatGPT, which Sam Altman helped bring into the mainstream in 2022. Users across the world have been experimenting with the chatbot to build, learn, and create things that once required significant time and expertise.

Last week, one such story grabbed global attention. Australian tech entrepreneur Paul Conyngham claimed he had developed a cancer vaccine for his dog, Rosie, using AI chatbots. According to his account, this was not just experimental. He says the treatment led to a 75 per cent reduction in the tumour.

Surprising? Absolutely. Confusing? Just as much.

Even Sam Altman thought so. Curious enough to understand how this was even possible, he decided to speak to Paul directly.

Sam Altman on speaking to Paul

After the meeting, Sam Altman took to X and described it as the coolest meeting he had that week.

He wrote, “The coolest meeting I had this week was with Paul, who used ChatGPT and other LLMs to create an mRNA vaccine protocol to save his dog Rosie. It is an amazing story.”

More from Tech

He also shared what Paul has to say, “The chat bots empowered me as an individual to act with the power of a research institute - planning, education, troubleshooting, compliance, and yes, real scientific design work in converting genomic data to a vaccine prescription and designing the treatment protocol around it. But they worked alongside humans at every step. The combination is what made it possible.”

This comes after Paul published a detailed account on X, outlining how he went from a concerned pet owner to someone navigating advanced cancer research.

Before diving into the science, it is worth pausing on what makes this story compelling. At its core, this is not just about AI or innovation. It is about urgency, persistence, and the willingness to explore every possible option when faced with a life-threatening diagnosis.

What stands out is not just the technology, but how it enabled an individual with no formal background in biology to engage with deeply complex systems and collaborate with experts along the way.

How AI helped build a personalised cancer vaccine for a dog?

AI did not create a cancer vaccine at the click of a button. In Paul’s case, it acted as an indispensable co-pilot, enabling him to navigate a deeply complex scientific process that would typically require years of formal training.

At the centre of this effort was data. After conducting full genome and RNA sequencing of his dog’s tumour, Paul was dealing with nearly 300GB of raw data. With no formal background in biology, he turned to AI systems like ChatGPT, Gemini and Grok to design a bioinformatics pipeline from scratch. These tools guided him on which specialised software to use, including alignment tools, variant callers, and annotation systems, and helped him interpret outputs at each step.

Through this process, hundreds of mutations were filtered down to a handful of viable neoantigen targets, the key building blocks for a personalised vaccine.

AI also played a crucial role in understanding the cancer itself, stated Paul. Using platforms like AlphaFold, Paul was able to model mutated proteins, particularly the c-KIT mutation driving the cancer, and analyse how these changes contributed to tumour growth. This kind of structural insight would have previously relied on advanced laboratory techniques.

When initial approaches such as ligand discovery and docking existing compounds proved unfeasible due to time and legal constraints, AI helped pivot the strategy towards immunotherapy. Specifically, it guided the design of a personalised mRNA neoantigen vaccine, shifting the focus from blocking cancer growth to training the immune system to recognise and destroy cancer cells.

Using AI, Paul identified seven key epitopes expressed in both DNA and RNA data. These were then assembled into an mRNA vaccine construct, optimised with linkers, adjuvants, and structural refinements to ensure stability and effectiveness. Multiple AI systems were used iteratively, one to generate the construct, another to refine it, and others to validate its properties.

Beyond vaccine design, AI helped shape the broader treatment protocol. It assisted in combining the vaccine with other therapies, including a tyrosine kinase inhibitor targeting the c-KIT mutation and a PD-1 checkpoint inhibitor to enhance immune response. AI tools were used extensively to model drug interactions, optimise timing, and plan a phased treatment rollout.

Equally important was AI’s role as a teacher and troubleshooter. It helped Paul understand complex concepts such as tumour immunology, neoantigens, and the tumour microenvironment, while also debugging technical issues in the pipeline and assisting with over 100 pages of ethics and regulatory documentation.

In essence, AI did not replace scientists or clinicians. However, it dramatically accelerated the process, reduced barriers to entry, and made it possible for a non-expert to contribute meaningfully to the development of a highly personalised cancer treatment.