Welcome to MedAI Roundup, highlighting the latest news and research in healthcare-related artificial intelligence each month.
And that’s just the beginning: ChatGPT parent company OpenAI said it has started development with Microsoft on the next-generation AI model, GPT-5. Notably, OpenAI CEO Sam Altman is back at the company after a tumultuous exit. (Financial Times and New York Times)
Science has updated its AI editorial policy to now permit some exceptions for using AI-generated images and videos in manuscripts.
STAT has developed a tracker of health systems and healthcare companies that are leading adoption of generative AI in medicine.
Yet another study has shown that generative AI models are capable of writing very convincing clinical notes that are hard to distinguish from physician-written notes, according to a paper published in NPJ Digital Medicine.
However, Nigam Shah, MBBS, PhD, Stanford Health Care’s chief data scientist, warned that AI-powered clinical support tools require carefully selected data to operate effectively and without bias, according to an interview with JAMA editor-in-chief Kirsten Bibbins-Domingo, MD, PhD.
Similarly, researchers at the Dana-Farber Cancer Institute issued a warning about the ethical implications of using patient-facing AI tools in a study published in JCO Oncology Practice.
The American Medical Association’s board also recently approved the association’s principles for the development and use of AI in healthcare, which focus primarily on transparency, oversight, and data privacy and security.
Still, only 13% of medical group leaders reported AI as their top tech priority, according to an MGMA poll of 424 healthcare leaders. Ensuring usability of electronic medical records was the top tech priority, at 35% of respondents.
The Department of Veterans Affairs launched a new $1 million competition to encourage researchers to create AI-enabled tools that can address burnout among healthcare workers.