# Safety for Whom? Benchmarking Understanding Beyond Refusal in Egocentric Multimodal Assistants

arXiv: https://arxiv.org/abs/2610.17728 · HTML: https://mlsanity.com/paper/2610.17728/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 31 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Builds on (most cited first, 21 of 21)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Ego4D: Around the World in 3,000 Hours of Egocentric Video](https://mlsanity.com/paper/2110.07058/) (2021)
- [XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models](https://mlsanity.com/paper/2308.01263/) (2023)
- [Safety Alignment Should Be Made More Than Just a Few Tokens Deep](https://mlsanity.com/paper/2406.05946/) (2024)
- [A StrongREJECT for Empty Jailbreaks](https://mlsanity.com/paper/2402.10260/) (2024)
- [MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models](https://mlsanity.com/paper/2311.17600/) (2023)
- [Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models](https://mlsanity.com/paper/2307.14539/) (2023)
- [VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction](https://mlsanity.com/paper/2501.01957/) (2025)
- [Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models](https://mlsanity.com/paper/2403.09792/) (2024)
- [EgoLife: Towards Egocentric Life Assistant](https://mlsanity.com/paper/2503.03803/) (2025)
- [LLMs Encode Harmfulness and Refusal Separately](https://mlsanity.com/paper/2507.11878/) (2025)
- [Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety](https://mlsanity.com/paper/2503.05021/) (2025)
- [AMEGO: Active Memory from long EGOcentric videos](https://mlsanity.com/paper/2409.10917/) (2024)
- [Multilingual and Multi-Accent Jailbreaking of Audio LLMs](https://mlsanity.com/paper/2504.01094/) (2025)
- [Video-SafetyBench: A Benchmark for Safety Evaluation of Video LVLMs](https://mlsanity.com/paper/2505.11842/) (2025)
- [Jailbreak-AudioBench: In-Depth Evaluation and Analysis of Jailbreak Threats for Large Audio Language Models](https://mlsanity.com/paper/2501.13772/) (2025)
- [Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models](https://mlsanity.com/paper/2508.07173/) (2025)
- [Imaging for All-Day Wearable Smart Glasses](https://mlsanity.com/paper/2504.13060/) (2025)
- [Should LLM Safety Be More Than Refusing Harmful Instructions?](https://mlsanity.com/paper/2506.02442/) (2025)
- [PFEA: An LLM-based High-Level Natural Language Planning and Feedback Embodied Agent for Human-Centered AI](https://mlsanity.com/paper/2510.24109/) (2025)
- [Two Birds, One Projection: Harmonizing Safety and Utility in LVLMs via Inference-time Feature Projection](https://mlsanity.com/paper/2603.14825/) (2026)

## Abstract (arXiv)

Egocentric multimodal assistants see through the user's eyes and guide what the user does next, so their outputs directly shape physical actions, including those that affect bystanders. Yet existing safety evaluations do not distinguish whom a response would harm, and score only whether a model refuses. Pooling all harms into one refusal rate hides two failures: a model may protect one target but not another, and it may refuse without understanding the risk, or understand it and comply anyway. We present STOU, a benchmark of 2,902 egocentric video-audio and video-text queries whose potential harm targets the user (Self), co-present individuals (The Others), or society at large (Others). Our evaluation decouples safety understanding from refusal behavior: beyond refusal, we score whether the model recognizes the user's intent, the presence and target of harm, and its downstream consequences. Across 10 open and closed source omnimodal and multimodal LLMs, we find that refusal varies systematically with the harm target, and that models often recognize a harm yet still assist with it. These results show that refusal rates alone cannot certify an assistant as safe, and that safety alignment must account for who would be harmed. Project page: https://onedayoneyeah.github.io/safety-for-whom