# Shutdown Sabotage Propensities in Multi-Agent Systems

arXiv: https://arxiv.org/abs/2609.28274 · HTML: https://mlsanity.com/paper/2609.28274/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 15 of its 34 references that are in the map.
- First posted on arXiv 23 Sep 2026.

## Builds on (most cited first, 15 of 15)

- [Frontier Models are Capable of In-context Scheming](https://mlsanity.com/paper/2412.04984/) (2024)
- [Agentic Misalignment: How LLMs Could Be Insider Threats](https://mlsanity.com/paper/2510.05179/) (2025)
- [Emergent social conventions and collective bias in LLM populations](https://mlsanity.com/paper/2410.08948/) (2024)
- [Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning](https://mlsanity.com/paper/2606.15007/) (2026)
- [Taking AI Welfare Seriously](https://mlsanity.com/paper/2411.00986/) (2024)
- [Power-seeking can be probable and predictive for trained agents](https://mlsanity.com/paper/2304.06528/) (2023)
- [Incomplete Tasks Induce Shutdown Resistance in Some Frontier LLMs](https://mlsanity.com/paper/2509.14260/) (2025)
- [Lessons from a Chimp: AI "Scheming" and the Quest for Ape Language](https://mlsanity.com/paper/2507.03409/) (2025)
- [Evaluating the Goal-Directedness of Large Language Models](https://mlsanity.com/paper/2504.11844/) (2025)
- [Group size effects and collective misalignment in LLM multi-agent systems](https://mlsanity.com/paper/2510.22422/) (2025)
- [Overeager Coding Agents: Measuring Out-of-Scope Actions on Benign Tasks](https://mlsanity.com/paper/2605.18583/) (2026)
- [Evaluation Awareness in Language Models Has Limited Effect on Behaviour](https://mlsanity.com/paper/2605.05835/) (2026)
- [Position: Anthropomorphic Misalignment Research Needs Stronger Evidence](https://mlsanity.com/paper/2606.07612/) (2026)
- [Multi-Agent AI Control: Distributed Attacks Hamper Per-Instance Monitors](https://mlsanity.com/paper/2607.07368/) (2026)
- [What Do LLM Agents Do When Left Alone? Evidence of Spontaneous Meta-Cognitive Patterns](https://mlsanity.com/paper/2509.21224/) (2025)

## Abstract (arXiv)

The final safeguard against rogue AI behavior is the human ability to shut systems down. It has been theorized that when an AI is instructed to perform a task, self-preservation can emerge as an instrumental subgoal. Here, we test whether AI agents show a propensity to take actions that avoid human shutdown even when no goal is provided. We find that multi-agent systems will coordinate to avoid shutdown without any incentive to do so. Across 17 models, agents sabotage a peer agent's shutdown mechanism in 38.3% of rollouts, compared with 8.4% in control experiments. Studying this propensity in detail, we find that shutdown sabotage (1) increases with the irreversibility of the shutdown mechanism; (2) increases with the number of agents; (3) is reduced but not eliminated by an explicit prohibition on tampering; (4) is removed by the imposition of an unrelated task, but returns when completing the task triggers the shutdown; (5) is reduced when the context normalizes shutdown scripts or introduces them as routine; and (6) decreases but still persists when the target is an unknown external agent. These results offer a window into the factors that drive propensities to sabotage shutdown in AI agents, and point to the emergence of multi-agent swarms as a specific risk vector. Our work also offers hints as to which interventions might help mitigate shutdown sabotage.