This report maps the transition from human-level AGI to artificial general superintelligence (ASI), defined as a system more cognitively capable than large organisations of humans. It identifies four potential pathways: scaling existing AGI, algorithmic paradigm shifts, , and ASI emerging from large-scale multi-agent collectives. These pathways are not mutually exclusive and may operate in parallel.
AI progress may continue past AGI toward superintelligence, not stop
Google DeepMind authors map four pathways from human-level AGI to ASI and identify bottlenecks, open research questions, and societal implications.
Big Tech
Tim Genewein · Matija Franklin · Alexander Lerchner · Laurent Orseau · Samuel Albanie · Adam Bales · +8 more
Google DeepMind · University of Waterloo · Australian National University · University College London
Research Digest··1 min read
In one line
Post-AGI progress to artificial superintelligence could occur through scaling, paradigm shifts, recursive improvement, and multi-agent collectives, with uncertain bottlenecks.
Before you read this
Two ideas this paper builds onrecursive self-improvementAn AI system repeatedly uses its own outputs to enhance its capabilities, creating a feedback loop for autonomous improvement.49 papers →multi-agent coordinationThe collaborative operation of multiple AI agents to solve tasks more effectively or with greater capability.77 papers →
Read it as
Genewein et al. present a report examining how AI might develop beyond human-level AGI to artificial superintelligence (ASI).
The authors, primarily from Google DeepMind, produced a structured investigation into the post-AGI trajectory of machine intelligence.
The Daily Edition
The front page in your inbox each weekday morning: the lead, the picks, and why each one earned its place.