In a recent demonstration, digital pathologist Dr. Aleksandra Zuraw tested two advanced AI models—GPT-5 and Gemini 2.5 Pro—on their ability to read and interpret pathology slides. The results highlight the growing collaboration between pathologists and artificial intelligence in the medical field.
"Both models showed impressive capabilities in identifying key features on the slides," said Dr. Zuraw. "However, they also had limitations, reminding us that AI is a tool to augment, not replace, expert human judgment."
The video, shared on the Digital Pathology Place channel, showcased how these large language models can assist in analyzing digital biomarkers, potentially speeding up diagnoses and reducing human error. Dr. Zuraw emphasized that while AI can flag abnormalities, final interpretations must be verified by trained pathologists.
This development comes as the healthcare industry increasingly adopts digital pathology and AI-driven diagnostics. The collaboration between pathologists and AI is expected to improve accuracy and efficiency, particularly in areas like cancer detection and rare disease identification.
For those interested in learning more, Dr. Zuraw offers a free eBook on digital pathology and courses on the subject through her website.