On December 3rd the Government Accountability Office (GAO) and the National Academy of Medicine (NAM) released a report on artificial intelligence (AI) tools’ potential applications in clinical and administrative health settings. This somewhat unusual collaboration addresses some of the many paths forward for the use of AI in health and healthcare. One key thing that caught our attention was the need for transparency on datasets used in algorithm development to understand their applicability and potential for bias that can evolve from the use of AI tools in healthcare. We urge that additional research on frameworks for AI design emerge and that this renews the push for open data sources. Increased data access will allow AI tools to be representative of diverse populations across a myriad of issues in population health and individual medical decision making.  The recent release of HHS COVID hospital data is an example of needed transparency and the urgency to get the data right.   

Read the full report:

GAO National Academy of Medicine Artificial Intelligence in Health Care

Link to the HHS hospital level data:

https://www.hhs.gov/about/news/2020/12/07/hhs-publishes-covid-19-hospital-facility-level-data.html 

Leave a Reply

Your email address will not be published. Required fields are marked *