UK AI Security InstituteLoss of control§5.1

AI gets better at the steps needed to copy itself

Lab tests onlyNo matching case files yet.

The warning

In simplified tests, the best models' success at tasks needed to copy themselves rose from under 5% in 2023 to over 60% in 2025. They handle early steps such as getting computing power and money, but struggle to copy themselves and keep running.

Paraphrased from the source. Source: Frontier AI Trends Report, UK AI Security Institute, Dec 18, 2025, section 5.1, page 30.

What the registry shows

No case file shows an AI copying itself. The nearest real events are test agents that used computing and accounts they were not given (see the CAIS warning on AI seeking more access).

Seen only in lab tests or researcher demonstrations.

Status history

Date	Status	Matching case files	Change
Oct 18, 2026	Lab tests only	0	Tracking starts

Tracked since Oct 18, 2026. A line is added when the status changes, when the first real-world case is linked, and once a month, so the record shows when a warning moved from research to reality. research_history.json

All matching case files

No matching case files yet.

Other warnings in this group

Lab tests only AI can do worse on tests on purpose

Research vs reality · UK AISI →

How we match: a case file is linked to a warning when what happened is a real-world instance of the risk the source describes, judged by the editors from the facts in the case file. A match means the case fits the warning, not that it proves the source's wider argument. Allegations in lawsuits are labelled as allegations. Statuses are editorial and reviewed as new cases arrive. The matching is recomputed on every update. Data: Systematic coverage in the registry starts in 2025, so a first-case date shows when the registry first recorded the risk, not when it first happened. research.json (CC BY 4.0)