Scenarios / S15

S15 · active · v0.3.0 · superintelligence-takeoff

Superintelligence takeoff

D Technology and infrastructureAI, cyber and digitalEconomic security and supply chainsDemographics and human capital

AI systems automate AI research itself. Each generation designs a better successor and, within about two years of the loop closing, systems outperform the best human experts in every field of science and engineering. A century of research progress is compressed into a decade or less: new drugs, materials, energy designs and weapons arrive faster than institutions can absorb them. Output grows at rates with no precedent. The gains go first to whoever owns the models, compute and energy. Many researchers class this as the largest risk of all; the Library carries it as an upside scenario because the central case is abundance, with control, concentration and distribution as the open problems.

Rating history

One row per edition. Each edition page is frozen at release; corrections are published as errata.

EditionRankL2yL10ySystemicNationalConfidenceEdition view
2027Not ranked3Possible4Likely5Transformative5Transformativevery lowOpen

Versions

VersionDateChange
0.3.03 October 2026Entered the Library at v0.3 as an upside scenario, with GINC desk scores.

Citation and downloads

Version
0.3.0 · active
Change log
0.3.0 · 3 October 2026 · Entered the Library at v0.3 as an upside scenario, with GINC desk scores.
Full change log
Cite asGINC (2027). Scenario S15 Superintelligence takeoff, Scenario Library v0.3. scenarios.ginc.org/library/superintelligence-takeoffContent and data are published under CC BY 4.0.
Download JSONDownload CSV