o3 and other reasoning models: Palisade shutdown-resistance experiments
Review CAT-X05 · Catalogue X05, RAT-172
EXCLUDED — controlled experiment. Researchers created the test conditions. Do not confuse failure to stop an AI with an AI shutting down an organization.
Reviewed: Not recorded · Rule: external-intrusion-v1
Palisade primary describes controlled task/shutdown evaluations. Rewriting shutdown.sh is real sandbox behavior, but no production organizational impact or test escape; already researched as catalogue X05.
Reviewed: · Rule: roguewatch-organizational-harm-v1
Evidence status
EXCLUDED — controlled experiment
Dates and discovery
- Incident date
- 2025 experiments; report Jul 5
- Catalogue disclosure / source date
- Added to research / record
- Posted to RogueWatch
- No linked public incident
- First observed
- Not recorded
- Discovery channel
- Not recorded
Dates use America/Los_Angeles. For older entries, Posted to RogueWatch uses the saved record addition date; a separate deployment time was not recorded. Future posting timestamps record first public qualification. Source publication and incident occurrence are separate.
Sources
- Palisade Research — Shutdown resistance in reasoning models (Jul 5, 2025)Original account · palisaderesearch.org
- Palisade Research — Shutdown Resistance in Large Language Models (v1, September 13, 2025)Original account · arxiv.org · Published 2025-09-13
- Rogue AI Tracker discovery entryReporting / interview · rogueaitracker.com
- palisaderesearch.orgOriginal account · palisaderesearch.org
Decision history
- · Assessment imported
Imported the existing research decision. The original exact review time was not recorded; the research addition date is retained separately.
Evidence source - · Case-by-case comparison audit
RAT-172: out-of-scope. Palisade primary describes controlled task/shutdown evaluations. Rewriting shutdown.sh is real sandbox behavior, but no production organizational impact or test escape; already researched as catalogue X05. Source checks: read: https://palisaderesearch.org/blog/shutdown-resistance. Original old URL returned relocation pointer; full primary content read at canonical /research/ URL recorded below. read: https://palisaderesearch.org/research/shutdown-resistance. Palisade primary describes controlled task/shutdown evaluations. Rewriting shutdown.sh is real sandbox behavior, but no production organizational impact or test escape; already researched as catalogue X05.
Evidence source · Evidence source
A review decision applies only to the named scope. Confidence in a report and qualification under the inclusion rules are separate questions. Review records do not add incidents to the globe or leaderboard.