And some of what it’s learned is uncomfortable. New research shows that when tech companies cleaned up the obvious racism in AI, they didn’t eliminate its bias. They just trained it to conceal it better.
AI learned to “sound” polite while making the same discriminatory decisions about jobs, healthcare, and criminal sentencing.
Then, a reflection on the work that chooses you and the responsibility that comes with it.
Things Worth Passing On: The Wand Chooses the Wizard
—Sharon McMahon, Editor-in-Chief
P.S. We’re working on a series about healthcare in America for next week’s issue of The Preamble. We’re looking at insurance, access, costs, the whole system, and I need your stories. Send us a letter about what you’ve experienced or what you wish people understood.