# Contested concept: recursive self-improvement

_**What you are reading**: one contested-concept page in Deliberus's deliberation graph. Deliberus maps human reasoning — including claims that are contested, mistaken, or extreme — so that disagreement can be examined precisely. A claim's presence here is not an endorsement: the graph records that someone argued it, how it connects to other reasoning, and how far scrutiny has gotten. The metadata on this page describes the state of that scrutiny, not the truth of the claim._

---

- **Contestedness**: semantic
- **Diagnosis**: This semantic shift masks a disagreement about whether the 'intelligence explosion' has already begun (gradually) or is a discrete future shock.
- **Canonical**: `/concept/recursive self-improvement` (data: `/concept/recursive self-improvement.json`)

## Senses in use
- **AI systems assisting in the generation of their own code or efficiency improvements under human supervision.** — used in 6 claim(s)
- **The total automation of the AI research and development process, replacing human researchers entirely.** — used in 3 claim(s)

## Claims using this term
- [Public disclosures from Anthropic, OpenAI, and Google DeepMind regarding AI-assisted development provide evidence that recursive self-impro…](/claim/claim_17792015a947.md) _(sense: AI systems assisting in the generation of their own code or efficiency improvements under human supervision.)_
- [Fully automating artificial intelligence research and development (R&D) remains a significantly difficult technical challenge or 'tall orde…](/claim/claim_75ea8536e910.md) _(sense: The total automation of the AI research and development process, replacing human researchers entirely.)_
- [The OpenAI model GPT-5.3-Codex assisted in the process of building its own system.](/claim/claim_7e5ecfa6e0b9.md) _(sense: AI systems assisting in the generation of their own code or efficiency improvements under human supervision.)_
- [Recursive self-improvement in artificial intelligence is a present-day phenomenon.](/claim/claim_83e0a9a48235.md) _(sense: AI systems assisting in the generation of their own code or efficiency improvements under human supervision.)_
- [There is a 10% probability that artificial intelligence research and development will be fully automated by the end of 2026.](/claim/claim_9a24bb84b238.md) _(sense: The total automation of the AI research and development process, replacing human researchers entirely.)_
- [Recursive self-improvement of artificial intelligence is not a future phenomenon but is currently occurring in the broadest sense.](/claim/claim_b69256003a90.md) _(sense: AI systems assisting in the generation of their own code or efficiency improvements under human supervision.)_
- [Between 70% and 90% of the code used in developing future artificial intelligence models at the company Anthropic is currently written by i…](/claim/claim_b70e2d9d58ba.md) _(sense: AI systems assisting in the generation of their own code or efficiency improvements under human supervision.)_
- [A Google DeepMind artificial intelligence model improved the efficiency of its own training process.](/claim/claim_d8dd2478af28.md) _(sense: AI systems assisting in the generation of their own code or efficiency improvements under human supervision.)_
- [OpenAI is aiming to develop a meaningful, fully automated artificial intelligence researcher by March of 2028.](/claim/claim_e3c5b43c7601.md) _(sense: The total automation of the AI research and development process, replacing human researchers entirely.)_