NEJM AI Grand Rounds

The Double-Edged Sword of AI, with Dr. Ziad Obermeyer

4 snips
Jul 27, 2023
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Problem Formulation and Bias

  • Machine learning problem formulation is crucial for algorithmic bias.
  • Choosing the prediction target reflects societal biases and affects model generalizability.
ANECDOTE

Racial Bias in Healthcare Algorithm

  • Dr. Obermeyer's team analyzed a healthcare algorithm used by a large insurer.
  • They found racial bias due to the algorithm predicting cost, not actual health needs.
ADVICE

Navigating Algorithmic Bias Research

  • Remember that accessing algorithm scores is a privilege in healthcare research, unlike other fields.
  • Approach companies with a positive attitude, as most people want to do the right thing and fix biases.
Get the Snipd Podcast app to discover more snips from this episode
Get the app