What Is Superintelligence? ASI vs AGI, Examples, Risks
What superintelligence is, how ASI differs from AI and AGI, proposed paths, possible uses, the main risks, and why no system qualifies today.
Reviewed 4 Nov 2026
What is established about AI, and what is only claimed.
A reference on artificial intelligence, AGI and superintelligence. Every claim carries an evidence grade, every figure a date, every forecast a record you can check later.
- Articles
- 139
- Chapters
- 68
- Glossary terms
- 291
Every article was checked once against primary sources, between October and November 2026. Claims that could not be confirmed were softened or removed. How this works.
Discover
A random pick of articles and definitions from the encyclopedia.
- Volume VI · ArticleArtificial SuperintelligenceArtificial Superintelligence (ASI) refers to a hypothetical form of machine intelligence that greatly exceeds human cognitive performance across…Read the article
- Glossary · DefinitionCapabilityWhat an AI system can do when appropriately prompted, scaffolded or fine-tuned, as opposed to what it does by default. Safety evaluations…See the definition
- Volume XII · ArticleForecasting AI Progress: Methods and CalibrationForecasting AI progress is the practice of making explicit, checkable statements about future AI capabilities, resources, or impacts, and of judging…Read the article
- Glossary · DefinitionGoal misgeneralizationA failure in which a system keeps its capabilities in a new situation but pursues a different goal from the one its designers intended, because the…See the definition
- Volume II · ArticleThe Transformer and Foundation Model Era, 2017-2022The Transformer and Foundation Model Era is the period from the 2017 publication of the Transformer architecture to the public release of ChatGPT at…Read the article
- Glossary · DefinitionOmega PointTeilhard de Chardin's name for a final point toward which consciousness and complexity are said to converge. Frank Tipler later used the name for a…See the definition
- Volume III · ArticleGenerative Adversarial NetworksA generative adversarial network (GAN) is a framework for training a generative model in which two neural networks are trained against each other: a…Read the article
- Glossary · DefinitionSymbol grounding problemThe question of how symbols in a formal system acquire meaning beyond their relations to other symbols, posed by Stevan Harnad.See the definition
From the blog
News, updates and sourced explainers. All posts
What Would Superintelligence Need? Power, Compute, Capital
No ASI exists and no source says what it would need. What labs use today: power, clusters, data, capital and security, with dated, sourced figures.
AI vs AGI vs SI vs ASI: Definitions and Differences
How AI, AGI, superintelligence (SI) and ASI differ: definitions, why AGI has no agreed test, proposed pathways to ASI, examples and open questions.
Agents Don't Need to Be Smarter. They Need Permission Slips
2023 benchmarks: agents finished far fewer realistic tasks than people. Approval gates are a practical control, not proof discipline beats intelligence.
Volumes and chapters
Pick a chapter. Chapters with no article yet show their planned scope.
Foundations of Intelligence
Establish the concepts, definitions, scientific disciplines, and intellectual traditions required to understand artificial intelligence.
- 1What Is Intelligence?3 articles
- 2The Nature of Knowledge and Learning2 articles
- 3Mathematics and Logic6 articles
- 4The Brain and Biological Intelligence1 article
- 5Philosophy of Mind and Intelligence2 articles
14 articles · 5 chapters
History of Artificial Intelligence
Trace the development of artificial intelligence from its intellectual precursors to contemporary systems.
- 6Precursors to Artificial Intelligence1 article
- 7The Birth of Artificial Intelligence, 1940s–1950s1 article
- 8Early AI Development, 1950s–1960s3 articles
- 9Expert Systems and Early Commercialization, 1960s–1980s1 article
- 10AI Winters and Research Transitions1 article
- 11Statistical Machine Learning, 1980s–2000s2 articles
- 12Deep Learning, 2000s–20162 articles
- 13The Transformer and Foundation Model Era, 2017–20221 article
- 14The Contemporary AI Era, 2022 Onward1 article
13 articles · 9 chapters
Scientific & Technical Foundations
Explain how AI systems are constructed, trained, evaluated, and deployed.
- 15Programming and Computational Foundations2 articles
- 16Data and Information Systems1 article
- 17Machine Learning2 articles
- 18Deep Learning and Neural Networks11 articles
- 19Language and Generative Models7 articles
- 20AI Agents and Autonomous Systems3 articles
- 21AI Infrastructure and Hardware3 articles
29 articles · 7 chapters
The AI Capability Landscape
Establish a consistent framework for describing what AI systems can do, how those capabilities are measured, and how different forms of intelligence relate to one another.
- 22Categories and Types of AI2 articles
- 23Capabilities and Performance3 articles
- 24AI Evaluation and Measurement2 articles
- 25Comparing AI Systems2 articles
9 articles · 4 chapters
Artificial General Intelligence
Examine the meaning, scientific requirements, proposed pathways, and unresolved questions surrounding AGI.
- 26Defining AGI2 articles
- 27Scientific Approaches to AGI3 articles
- 28The Requirements for General Intelligence1 article
- 29AGI Timelines and Forecasting1 article
- 30Testing Claims of General Intelligence1 article
8 articles · 5 chapters
Artificial Superintelligence
Explore intelligence that substantially exceeds human capabilities, including the technical, strategic, philosophical, and societal questions that such systems could raise.
- 31Defining Superintelligence2 articles
- 32Potential Pathways to Superintelligence2 articles
- 33Intelligence Explosion Hypotheses1 article
- 34AI Alignment and Control1 article
- 35Existential Risk and Long-Term Futures1 article
- 36Human and Machine Coexistence1 article
8 articles · 6 chapters
Applications & Industries
Document how artificial intelligence is applied across the economy, sciences, public institutions, and everyday life.
- 37AI in Science and Research2 articles
- 38AI in Healthcare and Medicine1 article
- 39AI in Business and Economics1 article
- 40AI in Software and Computing1 article
- 41AI in Education and Knowledge Work1 article
- 42AI in Media and Creative Industries1 article
- 43AI in Industry and the Physical World1 article
- 44AI in Government and Society1 article
9 articles · 8 chapters
Safety, Ethics & Governance
Examine the risks, responsibilities, safeguards, and institutional structures associated with AI development and deployment.
- 45Technical Safety and Reliability6 articles
- 46AI Security1 article
- 47Ethics and Human Rights3 articles
- 48AI Law and Regulation4 articles
- 49Global AI Governance2 articles
16 articles · 5 chapters
The AI Economy
Explain the economic systems, organizations, resources, and incentives that shape AI development.
- 50The AI Industry1 article
- 51Economics of Artificial Intelligence2 articles
- 52Infrastructure, Resources, and Constraints1 article
- 53AI and Geopolitics1 article
- 54Measuring the AI Economy1 article
6 articles · 5 chapters
People, Organizations & Sources
Preserve the institutional and human history of AI.
- 55Pioneers and Influential Researchers12 articles
- 56AI Research Organizations3 articles
- 57AI Companies and Technology Providers1 article
- 58Foundational Publications and Research Papers1 article
- 59Primary Sources and Historical Records1 article
18 articles · 5 chapters
Philosophy & Consciousness
Examine the questions raised when artificial systems approach or potentially exceed human cognitive capabilities.
- 60Machine Consciousness1 article
- 61AI, Identity, and Personhood1 article
- 62Civilization and Advanced Intelligence1 article
- 63AI and the Cosmos1 article
4 articles · 4 chapters
Forecasts & Scenarios
Transform the encyclopedia from a record of existing knowledge into a disciplined system for tracking and evaluating possible futures.
- 64Forecasting AI Progress1 article
- 65Near-Term Outlook1 article
- 66Medium- and Long-Term Scenarios1 article
- 67Signals, Indicators, and Inflection Points1 article
- 68The Uncertainties That Matter Most1 article
5 articles · 5 chapters
Reference and method
Look things up, or see how the encyclopedia is built.
A–Z
Glossary
291 terms defined and cross-linked.
A–ZIndex
1508 entries across articles, chapters and terms.
ReferenceAI/SI Factbook
This is a table-based reference of verifiable facts about artificial intelligence: historical milestones, landmark benchmark and competition results, standard definitions and measurement conventions, major awards, and key…
ReferenceForecast Register
A living register of 26 documented public forecasts by other parties about AI capability, adoption, AGI and ASI timing, and risk. Each record uses the field set of docs/EVIDENCEFRAMEWORK.md section 2, plus the original…
ReferenceResearch Library
An index of 120 foundational works (books, seminal papers and major reports) that support the encyclopedia, grouped by subject. Each row records who wrote the work, when and where it appeared, a persistent identifier, its…
ReferenceAI/SI Model Registry
A living register of 55 historically or technically significant AI models and systems, from the 1958 perceptron to systems announced in September and October 2026. Each record states developer, source date, architecture family…
ReferenceAI/SI Timeline
This is a dated chronology of artificial intelligence, from the logic and calculating machines of antiquity and the 17th century to October 2026. It covers foundational theory, papers, algorithms, hardware, model releases,…
MethodEvidential & Epistemic Framework
Every substantive statement regarding AI capabilities, mathematical proofs, historical milestones, and future outlooks must be anchored to one of six evidential categories:
MethodSource Policy & Citation Standards
Articles must be constructed from an identifiable body of authoritative evidence. Contributors must never draft narrative assertions from unassisted recall and then attempt to match citations after the fact.
MethodEditorial Checklist, Linking Conventions & Status Definitions
Every article submitted or updated must be validated against the project's 15 Quality Dimensions (specified in docs/source/ENCYCLOPEDIABLUEPRINT.md, Section 19):