Researchers Map Possible Paths From AGI to Superintelligence
A new analysis argues that as artificial general intelligence (AGI) nears feasibility, researchers are shifting focus to what could follow: artificial superintelligence (ASI), a system that would consistently outperform large groups of human experts across nearly all cognitive tasks. The piece frames general intelligence as a continuum, with AGI as a midpoint and ASI as a far more capable stage requiring rigorous tracking methods to measure progress.
GoKawiil's interpretation of the reporting above, not reported fact.
This reframing suggests that AI governance and safety discussions may need to extend beyond human-level benchmarks toward preparing for systems whose impact on science and the economy could be far larger and faster than AGI's. It implies that current evaluation tools may be insufficient, which could push researchers and policymakers to develop new metrics and oversight mechanisms ahead of ASI's emergence.
- The discussion treats AGI and ASI as points on a single continuum of general intelligence.
- ASI is defined as consistently outperforming large groups of human experts across nearly all tasks, not just narrow domains.
- The analysis calls for better methods to track AI progress as development moves toward superintelligence.
Source: institute.deepmind.com — Tim Genewein, 2026-11-20
Published there as: “What Comes After AGI?”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.