Back
Recursive Self-Improvement (Wikipedia)
referenceCredibility Rating
3/5
Good(3)Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.
Rating inherited from publication venue: Wikipedia
A general-audience Wikipedia overview of recursive self-improvement; useful as an introductory reference but flagged for lacking citations and potentially containing original research as of February 2026.
Metadata
Importance: 45/100wiki pagereference
Summary
This Wikipedia article explains recursive self-improvement (RSI), the theoretical process by which an AGI system rewrites its own code to enhance capabilities, potentially triggering an intelligence explosion leading to superintelligence. It covers the 'seed AI' concept coined by Eliezer Yudkowsky, architectural requirements, and associated AI safety and ethics concerns.
Key Points
- •RSI describes AGI systems that rewrite their own code, potentially causing an intelligence explosion and leading to superintelligence.
- •The 'seed AI' concept, coined by Eliezer Yudkowsky, refers to an initial architecture capable of bootstrapping recursive self-improvement.
- •A hypothetical seed improver uses an advanced LLM with programming, planning, testing, and code-execution capabilities to iteratively improve itself.
- •RSI raises significant AI safety concerns, as self-modifying systems may evolve unpredictably and exceed human control or understanding.
- •The article notes Wikipedia quality concerns: portions are unsourced or rely on primary research, limiting its reliability as a reference.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| Self-Improvement and Recursive Enhancement | Capability | 69.0 |
Cached Content Preview
HTTP 200Fetched Mar 20, 202617 KB
# Recursive self-improvement
Recursive self-improvement
| | |
| --- | --- |
|  | This article **may contain [original research](https://en.wikipedia.org/wiki/Wikipedia:No_original_research "Wikipedia:No original research")**. Significant portion of text are unsourced or sourced to primary research papers. Please [improve it](https://en.wikipedia.org/w/index.php?title=Recursive_self-improvement&action=edit) by [verifying](https://en.wikipedia.org/wiki/Wikipedia:Verifiability "Wikipedia:Verifiability") the claims made and adding [inline citations](https://en.wikipedia.org/wiki/Wikipedia:Citing_sources#Inline_citations "Wikipedia:Citing sources"). Statements consisting only of original research should be removed._(February 2026)__([Learn how and when to remove this message](https://en.wikipedia.org/wiki/Help:Maintenance_template_removal "Help:Maintenance template removal"))_ |
Concept in artificial intelligence
Not to be confused with [Personal development](https://en.wikipedia.org/wiki/Personal_development "Personal development").
**Recursive self-improvement** ( **RSI**) is a process in which early [artificial general intelligence](https://en.wikipedia.org/wiki/Artificial_general_intelligence "Artificial general intelligence") (AGI) systems rewrite their own computer code, causing an [intelligence explosion](https://en.wikipedia.org/wiki/Intelligence_explosion "Intelligence explosion") resulting from enhancing their own capabilities and intellectual capacity, theoretically resulting in [superintelligence](https://en.wikipedia.org/wiki/Superintelligence "Superintelligence").[\[1\]](https://en.wikipedia.org/wiki/Recursive_self-improvement#cite_note-1)[\[2\]](https://en.wikipedia.org/wiki/Recursive_self-improvement#cite_note-2)
The development of recursive self-improvement raises significant [ethical](https://en.wikipedia.org/wiki/Ethics_of_artificial_intelligence "Ethics of artificial intelligence") and [safety](https://en.wikipedia.org/wiki/AI_safety "AI safety") concerns, as such systems may evolve in unforeseen ways and could potentially surpass human control or understanding.[\[3\]](https://en.wikipedia.org/wiki/Recursive_self-improvement#cite_note-3)
## Seed improver
The concept of a "seed improver" architecture is a foundational framework that equips an AGI system with the initial capabilities required for recursive self-improvement. This might come in many forms or variations.
The term "Seed AI" was coined by [Eliezer Yudkowsky](https://en.wikipedia.org/wiki/Eliezer_Yudkowsky "Eliezer Yudkowsky").[\[4\]](https://en.wikipedia.org/wiki/Recursive_self-improvement#cite_note-4)
### Hypothetical example
The concept begins with a hypothetical "seed improver", an initial code-base developed by human engineers that equips an advanced future [large language model](https://en.wikipedia.org/wiki/Large_language_model "Large languag
... (truncated, 17 KB total)Resource ID:
42900576efb2f3c1 | Stable ID: YTAwNTNjYm