One of the systemic risks the GPAI Code of Practice specifies: risks from humans losing the ability to reliably direct, modify or shut down a model, which may emerge from misalignment, self-replication, deception, resistance to goal modification or power-seeking 1. Signatories assess it for models with systemic risk; a deployer of agents asks how autonomy and tool use were evaluated.
- Developed in
- ch. 23, EU AI Act hooks for agents
ch. 08, GPAI Code of Practice - Chapters
- ch. 08, Regulatory Map · ch. 23, AI Agents
- Source
- 1 numbered reference, listed below
Where it is used
2 chapters of the Body of Knowledge use the term. Each link opens the first section that does.
- 08 · Regulatory Map GPAI Code of Practice 3 mentions
- 23 · AI Agents EU AI Act hooks for agents 1 mention
Related terms
Sources
- [1] General-Purpose AI Code of Practice, Safety and Security chapter (Commitment 9, serious incident reporting, Measure 9.2; Appendix 1.4 specified systemic risks, incl. loss of control). European Commission. 2025-07-10. https://ec.europa.eu/newsroom/dae/redirection/document/118119 (verified: primary)
Definitions of legal terms paraphrase the cited text, which governs. Dated statements are as of .