Back to Blog
AI & Machine Learning Published October 11, 2026 18 min read 77 views

Recursive Self-Learning in AI: Breakthroughs and Limits (2026)

AI is starting to improve its own tools, training and research workflows. Explore the science behind recursive self-learning, what works, and where it fails.

By Ghulam

Editorial illustration of a connected AI network surrounded by circular feedback paths, beside the words Recursive self-learning on a dark teal background.
Keep default if fields are optional; optional caption: "Recursive self-learning involves more than repeated model updates: the system may also change how future improvements are discovered." · Image credit: Asagus

Sources & References

  1. Schmidhuber, J. (2003; rev. 2006). Gödel Machines: Self-Referential Universal Problem Solvers Making Provably Optimal Self-Improvements
    Research Paper
  2. Zelikman, E. et al. (2022). STaR: Bootstrapping Reasoning With Reasoning
    Research Paper
  3. Madaan, A. et al. (2023). Self-Refine: Iterative Refinement with Self-Feedback .
    Research Paper
  4. Shinn, N. et al. (2023). Reflexion: Language Agents with Verbal Reinforcement Learning
    Research Paper
  5. Fernando, C. et al. (2024). Promptbreeder: Self-Referential Self-Improvement via Prompt Evolution
    Research Paper
  6. Zhang, J. et al. (2025). Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents
    Research Paper
  7. Zweiger, A. et al. (2025). Self-Adapting Language Models (SEAL)
    Research Paper
  8. Novikov, A. et al. (2025). AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery .
    Research Paper
  9. Lu, C. et al. (2026). Towards End-to-End Automation of AI Research
    Research Paper
  10. Srikanth, D. et al. (2026). Recursive Self-Improvement of AI Research Agents (AIDE²)
    Research Paper
  11. Xia, P. et al. (2026). RRSI: Regularized Recursive Self-Improvement of Agent Harnesses
    Research Paper
  12. Lee, H. et al. (2026). MASS: Recursive Self-Improvement through Multi-Agent Self-Supervision
    Research Paper
  13. Meng, F. et al. (2026). RSIBench-Data: Benchmarking Data-Centric Research for Recursive Self
    Research Paper
  14. Song, X., Cai, Z. & Zhao, L. (2026). When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs
    Research Paper
  15. Shumailov, I. et al. (2024). AI Models Collapse When Trained on Recursively Generated Data
    Research Paper
  16. Lee, H. et al. (2026). Recursive Harness Self-Improvement
    Research Paper
  17. Yin, S. et al. (2026). Recursive Self-Improvement via On-Policy Distillation for Reasoning
    Research Paper
  18. Wang, J. (2026). What Stops Recursive Self-Improvement in Robotics? Lessons from 123 Rounds of Agentic Skill Discovery
    Research Paper
  19. Cunningham, T. et al. (2026). The Economics of Recursive Self-Improvement
    Research Paper
  20. Chen, M. et al. (2026). Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loops
    Research Paper
  21. Zhao, A. et al. (2025). Absolute Zero: Reinforced Self-play Reasoning with Zero Data
    Research Paper
  22. Ho, S.-T., Liu, M. & Huang, F. (2026). When AI Reviews Train AI Reviewers: Scientific-Judgment Collapse and Mitigation
    Research Paper