Harvard School of Dental Medicine launched a student research initiative to create a unique software dedicated to detecting algorithmic bias in healthcare research. This program will analyze the working process of machine learning models that sometimes experience a lack of fairness due to various factors – from poor data providing to technical mistakes in the learning algorithm. These missteps sometimes may bring severe consequences, providing a false research result that may lead to the wrong practical outcomes.
The Generative Artificial Intelligence Summer Research Program is a "spin-off" of the Generative AI Research Program for young specialists interested in cross-disciplinary collaboration aiming to improve the AI industry.
How does it work?
Algorithmic bias is a subject of significant concern and a critical factor explaining why rapid AI implementation in medicine is very far from reality. That's why the the initiative pays so much attention to this issue.
The research team is creating a large language model (chatbot) with an extensive program fairness knowledge base. This developing interactive tool will recognize and report cases of AI biases occurring during the Machine Learning process. The LLM will also directly communicate with a "biological user," addressing any concerns in algorithm work. Potentially, the team is planning to extend bias-detector specialization beyond the healthcare industry, for instance, to finance research that uses AI technology.
Who works on the project?
Dr. Hawazin Elani, an assistant professor of oral health policy and epidemiology at the Harvard institution, runs this project. This initiative is dedicated to attracting talented students – from undergraduates to PhDs – interested in cutting-edge digital technology research that can benefit the healthcare industry.
Read other articles about dental education on our Newsfeed:
