Artificial superintelligence (ASI)
Yapay süper zekâ (ASI)
D1 · Foundations, models and data
ASI denotes artificial intelligence that would substantially surpass the best human cognitive abilities across virtually all relevant domains.
Review status: 2026-12-07
Technical explanation
The emphasis is broad superiority, not exceptional skill in one task.
Conceptual boundaries
Human-level general capability alone does not establish ASI.
Provider-neutral example
Counterexample: Bostrom excludes Deep Blue from general superintelligence because its competence is confined to chess.
Limitations
This is a conceptual capability category, not an arrival forecast or a claim about an existing product.
Related concepts
Atomic claims and evidence
1.1Nick Bostrom, Ethical Issues in Advanced Artificial Intelligence (2003, revised online version)
- Source
- Nick Bostrom, Ethical Issues in Advanced Artificial Intelligence (2003, revised online version)
- Source role
- Authoritative source
- Exact locator
- Section 1, Introduction, opening two paragraphs: broad superiority and the Deep Blue counterexample
- Supported claim
- Bostrom defines superintelligence by substantial superiority to the best human minds across practically every field.
- Last verification
- Review due
- Scope limitation
- His general-superintelligence definition is implementation-neutral; this entry concerns artificial systems.
2.1Morris et al., Position: Levels of AGI for Operationalizing Progress on the Path to AGI, arXiv v5
- Source
- Morris et al., Position: Levels of AGI for Operationalizing Progress on the Path to AGI, arXiv v5
- Source role
- Authoritative source
- Exact locator
- Section 4, Table 1 and final ASI paragraphs: narrow versus general superhuman capability
- Supported claim
- Levels of AGI distinguishes superhuman narrow AI from artificial superintelligence.
- Last verification
- Review due
- Scope limitation
- A proposed taxonomy, not a claim that ASI exists.