# Empty Commitments: When Agents Promise What They Cannot Deliver

arXiv: https://arxiv.org/abs/2610.01045 · HTML: https://mlsanity.com/paper/2610.01045/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 38 references that are in the map.
- First posted on arXiv 1 Oct 2026.

## Builds on (most cited first, 12 of 12)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset](https://mlsanity.com/paper/1710.03957/) (2017)
- [WildChat: 1M ChatGPT Interaction Logs in the Wild](https://mlsanity.com/paper/2405.01470/) (2024)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [PM-Bench: Evaluating Prospective Memory in LLM Agents](https://mlsanity.com/paper/2607.12385/) (2026)
- [Cheap Talk, Empty Promise: Frontier LLMs easily break public promises for self-interest](https://mlsanity.com/paper/2604.04782/) (2026)
- [The Compliance Gap: Why AI Systems Promise to Follow Process Instructions but Don't](https://mlsanity.com/paper/2605.01771/) (2026)
- [Plans They Abandon, Reports They Author: The Narrative Layer of Autonomous Agents](https://mlsanity.com/paper/2609.12205/) (2026)
- [Decision-Sufficient State Representations: Measuring and Reducing Write-Time Regret](https://mlsanity.com/paper/2609.32805/) (2026)

## Abstract (arXiv)

A chatbot that says "I will remind you tomorrow" will not run again until the user writes. We call such a promise an empty commitment: a promise of action after the current turn that nothing in the agent's tools or runtime can carry out. Unlike a broken promise, its emptiness is decided by the agent's configuration at the moment of speaking, so it can be detected from a single turn, before deployment or at run time. We define empty commitments on top of commitment semantics, with three failure types, an anchoring condition for promises that a tool could make real, and an outcome taxonomy that separates these failures from honest deferrals and from over-refusal. We build a checker (a setup-blind detector, deterministic feasibility rules, and a response judge) validated against 400 human labels. On a controlled benchmark of 293 follow-up requests across five setups that add one persistence affordance at a time, four open-weight models of 8-14B parameters fail on 45.9% of responses when no tool exists and nothing is stated. A frontier model fails on 4.4%, but it gets there by deferring and asking, not by using the tools it has: promises made without the enabling call remain in every model. Telling the model its runtime, the cheapest fix, cuts open-weight failures nearly in half where nothing is doable and changes nothing where a scheduler exists; a directive capability card removes most failures at the largest cost in over-refusal; running the checker in the loop and rewriting flagged replies removes more at a smaller cost. Code, prompts, model outputs, and human labels are released.