An experiment using two leading AI models (Copilot and Gemini) to summarize 15 publications yielded contradictory and incomplete results. This demonstrates that relying on AI output without rigorous human verification can lead to dangerously misinformed conclusions in medical communications.

In a partnership with Kenya's Penda Health, OpenAI conducted the first randomized controlled trial of an LLM co-pilot for physicians. The study demonstrated a statistically significant improvement in diagnosis and treatment outcomes for patients whose doctors used the AI assistant. This provides crucial real-world evidence that AI can move beyond lab benchmarks to tangibly improve care.

AI's true power in science isn't autonomous discovery, but process compression. It acts as an expert guide, allowing motivated individuals to navigate complex fields like drug discovery and assemble workflows that once required multiple specialized teams, blurring the line between professional research and individual effort.

AI is poised to revolutionize evidence synthesis by automating the grueling, multi-year process of systematic reviews. The ultimate goal is to enable anyone to get an accurate, near-instantaneous summary of the entire body of research on a specific question, effectively creating meta-analysis on demand.

The "AI vs. Dog Cancer" story shows that current AI's power is not autonomous discovery, but its ability to act as a research assistant, enabling motivated non-experts to orchestrate complex scientific projects by finding and coordinating with human experts.

AI identified circulating tumor DNA (ctDNA) testing as a highly sensitive method for detecting cancer recurrence earlier than scans or symptoms. Despite skepticism from oncologists who deemed it unproven, the speaker plans to use it for proactive monitoring—a strategy he would not have known about otherwise.

The widespread use of AI for health queries is set to change doctor visits. Patients will increasingly arrive with AI-generated analyses of their lab results and symptoms, turning appointments into a three-way consultation between the patient, the doctor, and the AI's findings, potentially improving diagnostic efficiency.

The progress of AI in predicting cancer treatment is stalled not by algorithms, but by the data used to train them. Relying solely on static genetic data is insufficient. The critical missing piece is functional, contextual data showing how patient cells actually respond to drugs.

An ordinary citizen, Paul Cunningham, used off-the-shelf AI like ChatGPT and Google's AlphaFold to design a custom mRNA vaccine that shrank his dog's tumor by 75%. This demonstrates the revolutionary potential of AI to empower individuals to solve complex scientific problems once exclusive to specialized experts.