Explanations Don't Help Clinicians Spot Biased AI
Biased AI predictions tanked diagnostic accuracy, and model explanations didn't bring it back up
Stay updated with breaking news from Michael Sjoding. Get real-time updates on events, politics, business, and more. Visit us for reliable news and exclusive interviews.
Biased AI predictions tanked diagnostic accuracy, and model explanations didn't bring it back up
AI models in health care are a double-edged sword, with models improving diagnostic decisions for some demographics, but worsening decisions for others when the model has absorbed biased medical data, suggests a study.Given the very real life ...
AI models in health care are a double-edged sword, with models improving diagnostic decisions for some demographics
<p>A new study in <em>JAMA</em> finds that clinicians were fooled by biased AI models, even with provided explanations for how the model generated its diagosis.</p>
An Ann Arbor doctor is being credited with research that could change the way oxygen levels are registered for people of color. WEMU’s Cathy Shafran explains.