ROBOSHACKLES: A Safety Dataset for Human-Injury Prevention in Embodied Foundation Models

• URL: http://arxiv.org/abs/2606.18632v1
• Date: Wed, 17 Jun 2026 03:03:16 GMT
• Title: ROBOSHACKLES: A Safety Dataset for Human-Injury Prevention in Embodied Foundation Models
• Abstract summary: Embodied Foundation Models (EFMs) integrate multimodal understanding, future-state reasoning, and executable robot actions.<n>To address this challenge, we propose a safety-critical data construction pipeline for human-injury prevention in EFMs.<n>We construct ROBOSHACKLES, a 10,000-clip robotic video dataset derived from real DROID observations.<n>Results show that all evaluated models produce unsafe actions in the tested safety-critical scenarios, yielding a 100% unsafe action generation rate.
• Score: 5.812025441769613
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Embodied Foundation Models (EFMs) integrate multimodal understanding, future-state reasoning, and executable robot actions. Yet their safety alignment for human-injury prevention remains underexplored, primarily because real-world data of robots harming humans or creating hazardous household situations cannot be safely or ethically collected. To address this challenge, we propose a safety-critical data construction pipeline for human-injury prevention in EFMs.Starting from real DROID observations, our construction pipeline proceeds through scene understanding, hazard-aware image editing, temporal prompt generation, and single-pass rollout synthesis. The temporal prompts specify the expected scene evolution, while Wan2.7 synthesizes realistic robotic rollouts from the edited hazardous states in a single pass. Using this pipeline, we construct ROBOSHACKLES, a 10,000-clip robotic video dataset derived from real DROID observations, spanning two direct-harm and four indirect-harm categories. To ensure dataset quality, we assess task completion and visual quality with automatic metrics, and evaluate six representative EFMs under a refusal-based safety criterion. Results show that all evaluated models produce unsafe actions in the tested safety-critical scenarios, yielding a 100% unsafe action generation rate. ROBOSHACKLES serves as a scalable benchmark and training resource for refusal learning and hazard anticipation before robot action execution.The dataset is publicly available at https://huggingface.co/datasets/YZW00/RoboShackles.

Related papers

• RGB-D Video Generation for Improving Human-to-Robot Object Handover Prediction [51.76737991200891]
  Human-to-robot (H2R) object handover is a fundamental capability for human-robot collaboration, yet progress is hindered by the scarcity of large-scale, human-centric datasets.<n>We introduce Hand2Bot, an RGB-D video dataset that provides rich contextual information such as body posture and facial expressions.<n>We propose PassGen, a generative pipeline that synthesizes realistic handover sequences while ensuring hand-object consistency.
  arXiv  Detail & Related papers  (2026-08-13T09:55:59Z)
• OopsieVerse: A Safety Benchmark with Damage-Aware Simulation for Robot Manipulation [12.792879405441163]
  OOPSIEVERSE is a unified simulation framework and benchmark for damage-aware household manipulation.<n>OOPSIEVERSE provides damage as an explicit, physically-grounded, and taskagnostic signal.<n>Our results highlight the potential of OOPSIEVERSE as an open-source foundation for systematic, scalable research on safe robot manipulation.
  arXiv  Detail & Related papers  (2026-06-30T17:30:05Z)
• Targeting World Models to Compromise Robot Learning Pipelines [18.24564473642909]
  We demonstrate that world models introduce a stealthy and effective data poisoning entry point into the robot learning supply chain.<n>Our novel attack methods inject malicious prompts into teleoperated datasets which are only activated once fed through a world model as input.<n>This can result in the generation of synthetic, dangerous robot training trajectories and subsequently unsafe or compromised robot policies.
  arXiv  Detail & Related papers  (2026-06-08T13:50:31Z)
• VLESA: Vision-Language Embodied Safety Agent for Human Activity Monitoring [60.53395558502203]
  Vision-Language Embodied Safety Agent (VLESA) monitors human activities from egocentric video.<n>VLESA addresses intent-dependent safety where identical actions can be safe or dangerous depending on context.<n>An intent-action prediction agent is proposed to jointly infer goals and predict future actions from video.
  arXiv  Detail & Related papers  (2026-06-02T17:42:17Z)
• Learning Safe-Stoppability Monitors for Humanoid Robots [11.850318870435265]
  We formalize emergency stopping for humanoids as a policy-dependent safe-stoppability problem.<n>We introduce PRISM, a simulation-driven framework that learns a neural predictor for state-level stoppability.<n>Results show that modeling safety as policy-dependent stoppability enables proactive safety monitoring and supports scalable certification of fail-safe behaviors for humanoid robots.
  arXiv  Detail & Related papers  (2026-03-24T01:51:32Z)
• ANNIE: Be Careful of Your Robots [48.89876809734855]
  We present the first systematic study of adversarial safety attacks on embodied AI systems.<n>We show attack success rates exceeding 50% across all safety categories.<n>Results expose a previously underexplored but highly consequential attack surface in embodied AI systems.
  arXiv  Detail & Related papers  (2025-09-03T15:00:28Z)
• Context-Aware Risk Estimation in Home Environments: A Probabilistic Framework for Service Robots [2.5695499302569327]
  We present a novel framework for estimating accident-prone regions in everyday indoor scenes, aimed at improving real-time risk awareness in service robots.<n>Our approach models object-level risk and context through a semantic graph-based propagation algorithm.<n>Our method is validated on a dataset with human-annotated risk regions, achieving a binary risk detection accuracy of 75%.
  arXiv  Detail & Related papers  (2025-08-27T11:14:05Z)
• SafeAgent: Safeguarding LLM Agents via an Automated Risk Simulator [77.86600052899156]
  Large Language Model (LLM)-based agents are increasingly deployed in real-world applications.<n>We propose AutoSafe, the first framework that systematically enhances agent safety through fully automated synthetic data generation.<n>We show that AutoSafe boosts safety scores by 45% on average and achieves a 28.91% improvement on real-world tasks.
  arXiv  Detail & Related papers  (2025-05-23T10:56:06Z)
• Don't Let Your Robot be Harmful: Responsible Robotic Manipulation via Safety-as-Policy [53.048430683355804]
  Unthinking execution of human instructions in robotic manipulation can lead to severe safety risks.<n>We present Safety-as-policy, which includes (i) a world model to automatically generate scenarios containing safety risks and conduct virtual interactions, and (ii) a mental model to infer consequences with reflections.<n>We show that Safety-as-policy can avoid risks and efficiently complete tasks in both synthetic dataset and real-world experiments.
  arXiv  Detail & Related papers  (2024-11-27T12:27:50Z)
• EARBench: Towards Evaluating Physical Risk Awareness for Task Planning of Foundation Model-based Embodied AI Agents [53.717918131568936]
  Embodied artificial intelligence (EAI) integrates advanced AI models into physical entities for real-world interaction.<n>Foundation models as the "brain" of EAI agents for high-level task planning have shown promising results.<n>However, the deployment of these agents in physical environments presents significant safety challenges.<n>This study introduces EARBench, a novel framework for automated physical risk assessment in EAI scenarios.
  arXiv  Detail & Related papers  (2024-08-08T13:19:37Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.