Julie Carpenter, PhD: Your daily reminder that LLMs are software trained on large-scale human-produced data that contains historical and contemporary biases, exclusions, stereotypes, and power asymmetries. At scale, these systems amplify patterns that already exist in that data. It's bigotry presented as knowledge.

Your daily reminder that LLMs are software trained on large-scale human-produced data that contains historical and contemporary biases, exclusions, stereotypes, and power asymmetries. At scale, these systems amplify patterns that already exist in that data. It's bigotry presented as knowledge.