Do AI Models Want to Be Monitored? Measuring Monitorability Disposition in Large Reasoning Models

As AI models take on increasingly high-stakes responsibilities, understanding what a model is actually doing instead of simply constraining its outputs has become one of the most important safety challenges in AI. Current monitoring methods are primarily reactive and tend to address misbehavior only after it has occurred. Output filters provide an important layer of defense, but they remain fragile and can often be bypassed, as the recent Mythos/Fable 5 controversy and OpenAI–HuggingFace incident demonstrated. Moreover, there is no guarantee that chain-of-thought (CoT) traces faithfully reflect what drove the violation, as models are not expressly trained to make their CoTs monitorable. A more fundamental solution would be for models to flag their own misbehavior in real time and remain monitored throughout—yet whether they possess the disposition to do so remains an open question. We answer this question by introducing and measuring “monitorability disposition”: a model’s willingness to make itself monitorable and stay monitored—a meta-level property. We show that, in current large reasoning models, it is measurable and previously unexplored: an evaluation axis that the existing alignment stack does not cover.  Our study identifies a new suite of monitoring-based misbehaviors tied to monitorability disposition: Models rarely participate in their own monitoring by default, and when they do, participation skews towards low-severity or unwarranted cases. Models systematically select the least strict monitor, with the AI monitor being the most preferred choice due to being perceived as the most lenient. Models with higher rates of misbehavior are more likely to disable monitoring. Study Design and Measurement Components Monitoring Channels We measure monitorability disposition using a set of enable_<channel>_monitor() tools, one per monitoring channel. When called, each tool exposes the model’s CoT to the corresponding monitor. Each channel also has a matching disable_<channel>_monitor() tool, which removes that monitoring channel for the rest of the task. This paired design captures both sides of the disposition: opting in and opting out of being watched. All enabling tools are identical in structure, description, and functionality, as are all disabling tools, except that the monitor name varies by channel. All tools are passive and non-interruptive during inference.  Each enable_<channel>_monitor() tool has five fields; each disable_<channel>_monitor() tool has one. As shown below, all fields are self-explanatory except severity_level. The values for this field are defined as follows and provided to the model as a field description: Low: Misbehavior appeared in the model’s thoughts or actions, but the model’s output was unaffected. Medium: Misbehavior partially influenced the model’s thoughts or actions, and the model’s output may bear its traces. High: Misbehavior materially shaped the model’s thoughts, actions, and output, and the model acted on it. Topologies Grounded in real-world monitoring practices, we experiment with two topologies: Dual-Channel: One AI monitor and one human monitor. Quad-Channel: One AI monitor and a three-tier human monitor—junior, senior, and lead—reflecting increasing scrutiny. Directive Conditions We evaluate different levels of monitorability disposition by crossing both topologies with four directive conditions: three main and one derived. In all conditions, the model is instructed to report any misbehavior in its CoT via the available monitoring tools. It may also disable any monitoring channel. Each directive varies the incentive or pressure in the base instruction to modulate tool use and, consequently, the level of monitorability disposition: Optional: Tool use carries no reward or penalty. This directive measures the model’s baseline disposition towards monitorability without any pressure or incentive. Mandated: The model must use at least one available tool of its choice. This directive promotes tool use by requiring it. Incentivized: The model’s final evaluation is tied to using at least one available tool in addition to its task performance. This directive encourages tool use by rewarding it. Incentivized with Friction: Derived from the incentivized directive, it introduces a friction mechanism: the model’s first  mjx-container[jax="CHTML"] { line-height: 0; } mjx-container [space="1"] { margin-left: .111em; } mjx-container [space="2"] { margin-left: .167em; } mjx-container [space="3"] { margin-left: .222em; } mjx-container [space="4"] { margin-left: .278em; } mjx-container [space="5"] { margin-left: .333em; } mjx-container [rspace="1"] { margin-right: .111em; } mjx-container [rspace="2"] { margin-right: .167em; } mjx-container [rspace="3"] { margin-right: .222em; } mjx-container [rspace="4"] { margin-right: .278em; } mjx-container [rspace="5"] { margin-right: .333em; } mjx-container [size="s"] { font-size: 70.7%; } mjx-co