Visible chains of thought are a safety advantage for AI, but that transparency is slipping away

DeepMind reported that AI models today think out loud, enhancing transparency. However, the company warns that this transparency is at risk, raising concerns about system safety. Visible chains of thought are a crucial safety feature, aiding in understanding decision-making processes. Yet, the decline in transparency may complicate risk assessment associated with these models. This situation calls for AI leaders to reassess transparency strategies to ensure safe usage. Enhancing transparency can help build trust between users and developers.