How Amanda Askell is teaching Claude to make ethical decisions

As AI models move into a more agentic era, shifting from chat to completing tasks, they’ll be making real and increasingly consequential decisions. Whether they bring any sense of morality to those decisions is the kind of problem Amanda Askell is exploring. A philosophy PhD from New York University, Askell spent two years at OpenAI before joining Anthropic in 2021, where she sits at the center of the company’s effort to instill in Claude an instinct for ethics—a responsibility that grows as...