Reducing The Risk Of The Exponential Growth Of Automated Influence Operations

A new report from the Center for Security and Emerging Technology (CSET), OpenAI, and the Stanford Internet Observatory outlines the potential threats posed by language models if used to augment disinformation campaigns and introduces a framework for analyzing mitigation strategies. The report is the result of a partnership between these three organizations and a workshop involving 30 researchers, machine learning experts, and policy analysts.