This exercise aims to map the landscape of potential AI breakthroughs and their social consequences.
The speed and direction of AI development will likely be shaped by hardware and algorithmic breakthroughs, conceptual developments in areas such as neuroscience and statistics, and social changes across political, industrial, and public dimensions. Keeping track of these developments will help to prioritise subsequent research, as control methods and social ramifications of AI will depend on both the system’s architecture and the timeline for its arrival.
While standard horizon scanning methods are immediately applicable for these purposes, this project plan to reach beyond standard methodology in two ways. First, it actively seeks out areas with the greatest disagreement and attempt to resolve these uncertainties by engaging with AI ‘skeptics’. Second, the workshops will be highly cross-disciplinary, bringing together philosophers, historians, social scientists, ecologists, economists, engineers, legal scholars, and representatives from industry groups to discuss these issues. This horizon scanning exercises will inform CFI's other lines of research (particularly Value Alignment and Trust and Transparency).