Large language models, or LLMs, have shown great promise. In fact, many observers think that LLMs may soon scale up to create artificial general intelligence, or AGI—an AI that possesses human or superhuman capabilities. In a new paper, RAND researchers argue that, when planning for the potential emergence of AGI, policymakers should avoid over-optimizing for a single possible future, such as the aforementioned scaling of LLMs. Instead, they should develop a robust policy strategy based on multiple paths that could lead to AGI.

The National Wastewater Surveillance System was launched in September 2020. It connects independent, local wastewater efforts to form a national program to help track the spread of diseases, including COVID-19, influenza, respiratory syncytial virus (RSV), and mpox. The funding for this system will run out later this year. According to RAND experts, sustaining and further investing in wastewater surveillance is essential to preserving U.S. health and security: “We cannot afford to go back to the pre–COVID-19 days in which a disease emerges and we're flying blind.”