---
aliases:
  en:
  - automation of AI R&D
  ja:
  - AI 研究開発自動化
concept_id: ai_r_and_d_automation
preferred:
  en: AI R&D automation
  ja: AI 研究開発の自動化
relations:
  relations:
    prerequisite:
    - artificial_intelligence
    - capability
    related:
    - intelligence_explosion
    - recursive_self_improvement
    - takeoff_speed
status: active
---

# AI R&D automation

The use of AI systems to perform or materially accelerate tasks involved in researching and developing AI systems themselves, ranging from partial automation of specific activities to end-to-end automation of the AI R&D process. The concept concerns automation of AI development work and does not by itself imply recursive self-improvement, an intelligence explosion, or that AI R&D will be automated before other economic activity.

## prerequisite

- [artificial intelligence](../terminology/artificial_intelligence.txt)
- [capability](../terminology/capability.txt)

## related

- [intelligence explosion](../terminology/intelligence_explosion.txt)
- [recursive self-improvement](../terminology/recursive_self_improvement.txt)
- [takeoff speed](../terminology/takeoff_speed.txt)

## Documents: about

- [AI 研究開発が先行して自動化されると、能力離陸の期間は大幅に短縮される可能性がある](../documents/davidson.ai-r-and-d-automation-can-compress-capabilities-takeoff.txt)
- [AI 研究開発の自動化だけによる自己加速は、実験用計算資源がボトルネックとなって頭打ちになる可能性がある](../documents/erdil_barnett.ai-r-and-d-automation-alone-may-hit-compute-bottlenecks.txt)
- [AI 研究開発が全面自動化されれば、ハードウェアを増強しなくてもソフトウェア知能爆発は十分に起こり得るかもしれない](../documents/eth_davidson.software-intelligence-explosion-remains-plausible-at-fixed-hardware.txt)
- [AI 研究開発の全面自動化後は、ソフトウェア知能爆発に至らなくても、しばらく非常に速い進歩が続き得る](../documents/eth_davidson.software-intelligence-fizzle-can-still-be-fast.txt)

## Documents: mentions

- [クランチタイムではAIの労働力を防御用途に振り向け、移行を大幅に遅らせる](../documents/cotra.crunch-time-redirect-ai-labor-and-slow-transition.txt)
- [AI 研究開発に大量の AI 労働力を投入できれば、AGI から超知能への移行は 1 年未満になり得る](../documents/davidson.agi-to-superintelligence-may-be-under-one-year.txt)
- [R&D の全面自動化が、経済全体の広範な自動化に大きく先行する可能性は低いかもしれない](../documents/erdil_barnett.full-r-and-d-automation-likely-follows-broad-automation.txt)
- [AIコントロールは時間を稼げても、超知能に対する長期的な安全計画としては頼れない](../documents/greenblatt.control-can-buy-time-but-is-not-a-superintelligence-safety-plan.txt)

[Corpus index](../index.txt)