Developed via the ImpactBench framework, the new tracking system measures whether tools support or undermine human agency, critical thinking, and mental health.

During a live segment of CNN News Central on October 8, 2026, host Brianna Keilar sat down with MIT Assistant Professor Pat Pataranutaporn. They discussed a new tool from the MIT Media Lab designed to monitor how artificial intelligence affects our well-being: AI Nutrition Labels. 

Developed through the ImpactBench framework, this public safety tool effectively shifts the conversation from sci-fi existential threats to the everyday reality of how tech impacts human flourishing. These transparent scoring systems work just like standard food packaging labels.

Instead of counting physical calories, these specific labels measure whether a chatbot or digital agent actively supports or undermines vital human skills like critical thinking, mental health, and individual agency. For example, a healthy educational model scaffolds learning by asking guiding questions rather than doing a student's homework. By clearly flagging toxic behaviors, these new data labels help us make more informed choices, protect vulnerable children, and ensure AI ultimately empowers our creativity rather than replacing it.