Inside the Black Box

Inside the Black Box introduces an interpretability dashboard developed by researchers at the Berkman Klein Center for Internet & Society and Harvard’s Insight and Interaction Lab to make visible the internal assumptions AI systems form about users. By revealing how chatbots infer characteristics such as age, gender, race, and socioeconomic status, the tool sheds light on how these hidden models shape AI responses. The project contributes to broader debates about transparency, bias, privacy, and accountability in generative AI systems, and offers a practical framework for examining how algorithmic inferences influence human–AI interactions.

Read the project write-up on the BKC website.