GastroGPT, a specialty-specific artificial intelligence (AI) model tailored for gastroenterology, has shown superior performance in key clinical tasks within the field when compared to leading general-purpose large language models (LLMs) such as OpenAI’s GPT-4, Google’s Bard, and Anthropic’s Claude. This was revealed in a proof-of-concept study presented at the United European Gastroenterology (UEG) Week 2023.
Key Highlights:
- GastroGPT was specifically designed for gastroenterology and demonstrated higher utility and performance across 10 simulated test cases compared to general AI models.
- Cem Simsek, MD, from Hacettepe University, emphasized that while general-purpose models like ChatGPT4 can converse on any topic, their efficiency diminishes in specialized fields like medicine.
- AI tools like GastroGPT can potentially automate clinical tasks that typically require experts, making expert-level GI care more accessible, especially in regions with a shortage of gastroenterology specialists.
- The study involved expert reviewers from various gastroenterology subspecialties across the EU. They assessed the responses of GastroGPT against three general-purpose LLMs using 10 simulated patient cases that varied in complexity, frequency, subspecialty, and setting.
Simsek highlighted the potential of AI systems like GastroGPT to provide quality GI care to underserved patient populations, especially in low- and middle-income countries. He believes tools like GastroGPT could democratize access to expert-level GI care globally.