Trustworthy recursive self-improvement for AI systems with safeguards
Trustworthy Recursive Self-Improvement in the Era of Large Models: Foundation, Risks, Evaluation and Open Challenges
Artificial IntelligenceCryptography and Security
Summary
AI systems are increasingly changing themselves based on experience, but making sure these changes actually improve the system is tricky. The authors provide a framework for understanding how AI can safely and reliably improve itself repeatedly, without losing important behaviors or security. They identify key areas to focus on, such as keeping improvements real, maintaining fairness and safety, protecting sensitive data, and enabling clear human control. Their survey also outlines open challenges and ways to evaluate these self-improvements over time.
What this means in practice
- •For ai safety engineers: Guide the design of AI systems that modify themselves while ensuring trustworthiness and human oversight remains effective.
- •For machine learning platform builders: Develop improvement processes that monitor and evaluate AI updates to prevent degradation or unintended behavior over repeated cycles.
A survey. It maps existing work.
Authors
Jiayue Pu, Ang Gao, Jinyi Bai, Qipeng Wang, Tongyu Wen, Jiahao Zhao, Keyu Qu, Junhao Zhang, Chenglei Shen, Zhongxiang Sun, Jun Xu, Xiao Zhang
Abstract
AI systems built around large models increasingly use experience and feedback to make persistent changes to models, prompts, memories, tools, and workflows. Recursive self-improvement (RSI) extends this process to inductive biases and to the mechanisms that govern future updates, which can themselves be revised and reused across improvement cycles. Yet self-modification alone does not guarantee trustworthy improvement. Reported gains may reflect flawed evaluations, previously acquired capabilities may deteriorate, and compliance with behavioral constraints may weaken. Compromised artifacts or sensitive information may also persist in successor systems. Changes to improvement mechanisms may invalidate earlier verification results and make human oversight less effective. This survey presents an operational definition and process framework for RSI and organizes the requirements for trustworthy RSI into four interconnected themes: genuine and sustained improvement; behavioral constraints, alignment, and fairness; system security and privacy; and transparency and controllability. We examine these themes from three perspectives: Improvement Validity, Constraint Preservation, and Effective Oversight. This analysis covers improvement diagnosis, proposal generation, proposal evaluation, and update implementation (S1--S4), including the recursive application of these stages to improvement mechanisms (RI). We review risks, safeguards, and evaluation evidence for individual updates, effects inherited by successor systems, and revisions to improvement mechanisms, while distinguishing direct RSI findings from evidence in related systems. Finally, we examine evaluation settings, identify requirements for independent and longitudinal assessment, and discuss applications and open research problems.