Review Artificial Superintelligence

Artificial superintelligence is a speculative concept for an AI system that would exceed human capability across most or all cognitive domains of interest.

ASI discussions concern a possible system beyond both current narrow AI and proposed human-level general intelligence. The term describes a scenario, not an established technology category.

The concept influences long-range safety, governance, and investment debates, but it should not be used to blur the evidence about systems that exist now.

A policy exercise considers how institutions might respond if a future system surpassed expert human performance across research, strategy, engineering, and persuasion.

ASI is not an observed next product release or a settled prediction. Timelines, feasibility, and even definitions remain uncertain.

ASI is a speculative beyond-human scenario, not present capability evidence.

* [[artificial-general-intelligence|Artificial General Intelligence]] * [[artificial-intelligence|Artificial Intelligence (AI)]] * [[scenario|Scenario]] * [[epistemic-humility|Epistemic Humility]] * [[alignment|Alignment]]

* [International AI Safety Report 2026](https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026) — internationally backed scientific assessment of advanced general-purpose AI capabilities and risks; useful context for keeping superintelligence claims explicitly hypothetical and uncertain. * [Kim et al., *The Road to Artificial SuperIntelligence: A Comprehensive Survey of Superalignment*](https://arxiv.org/abs/2412.16468) — research survey defining ASI as hypothetical and reviewing the associated alignment literature; included as a scholarly source, not evidence that ASI exists.

* Treat ASI as a [[scenario|Scenario]] and preserve [[epistemic-humility|Epistemic Humility]].