Startup aims to cure AI chatbots of groupthink

A startup is developing techniques to reduce the tendency of large language models to produce similar, predictable outputs. The approach aims to introduce more diversity in AI responses, addressing a growing concern about homogenization in chatbot behavior. Early tests show promise in breaking the 'groupthink groove' common among leading models.