A successful AI transition must ensure that people, organizations and institutions can still establish and trust facts, make sense of explanations, understand complex reality, exercise meaningful judgment and coordinate action. With a machine now in the middle of how information is gathered, interpreted and acted on, the infrastructure behind those capabilities has to be reconsidered from first principles.
Between December 2025 and May 2026, we brought around 140 people from technology, investment, policy, academia and media together in New York, San Francisco, Harvard and Copenhagen around one question: what needs to be built for an AI-mediated information ecosystem to serve society, and where should investment begin?
Participants worked from the premise that AI would fundamentally reorganize the information ecosystem. Each summit followed the same progression: identify what people and societies need from information; determine the functions required to meet those needs; then examine the products, services and infrastructure that could deliver them.
Across the four summits, a central tension emerged: some of the functions most valuable to society may be among the hardest to finance commercially. Resolving it will require builders, investors, philanthropists and public institutions to work from a shared understanding of the emerging system.
The resulting report maps those opportunities and dependencies. It is also the starting point for the Signals at Scale studio, where the next stage of inquiry is to build, test and learn. Read the report, or explore the value map.