Futures with Digital Minds: Expert Forecasts in 2025
arXiv Preprint Archive August 1, 2025 preprint
Study at a glance
AI-extracted from the abstract| Characteristics | Expert survey Qualitative |
|---|---|
| Sample size | 67 |
| Population | Experts in digital minds research, AI research, philosophy, forecasting, and related fields |
| Keywords | Cs.cy Subjective experience Consciousness Ai forecasting emergence Timeline Probability Expert survey Future of ai Ai ethics rights Societal discussions Welfare capacity |
| Key findings | Experts assign a median 90% probability that digital minds are possible in principle, 65% probability of creation by 2100, and 20% by 2030, with many anticipating rapid growth in welfare capacity to match billions of humans within a decade after the first digital mind. |
Abstract
This report presents findings from an expert survey on digital minds takeoff scenarios. The survey was conducted in early 2025 with 67 experts in digital minds research, AI research, philosophy, forecasting, and related fields. Participants provided probabilistic forecasts and qualitative reasoning on the development, characteristics, and societal impact of digital minds, that is, computer systems capable of subjective experience. Experts assigned high probability to digital minds being possible in principle (median 90%) and being created this century (65% by 2100), with a non-negligible probability of emergence by 2030 (20%). Many anticipated rapid growth in digital mind welfare capacity, with collective welfare capacity potentially matching that of billions of humans within a decade after the creation of the first digital mind. Participants also expected widespread claims from digital minds regarding their consciousness and rights, and predicted substantial societal disagreement over their existence and moral interests. Views diverged on whether digital mind welfare will be net positive or negative. These findings provide evidence that bears on the extent to which preparing the world for the potential arrival of digital minds should be a priority across domains such as research and governance. However, these findings should be interpreted cautiously in light of the potential for systematic overrepresentation of experts who deem digital minds particularly likely or important.