Latest / AI Ethics with Fexingo: Bias, Safety, and Responsible Artificial Intelligence
How AI Models Are Learning Bias From Your Job Descriptions
Episode 92 of AI Ethics with Fexingo explores a subtle but pervasive source of bias in artificial intelligence: the job descriptions companies use to train their hiring algorithms. Lucas and Luna unpack a 2025 study from the AI Now Institute that found AI models trained on real job postings from Fortune 500 companies consistently penalized resumes with non-traditional education paths, gaps in employment, and names associated with minority groups. The hosts walk through how this happens — from data collection to model deployment — and discuss what companies like IBM and Unilever are doing to…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.