{"ID":6138270,"CreatedAt":"2026-07-09T01:07:32.349475501Z","UpdatedAt":"2026-07-11T13:47:13.783731841Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.07420","arxiv_id":"2607.07420","title":"Initiation Safety: A Missing Dimension in Generalist-Robot Safety","abstract":"Safety for generalist robots is usually discussed in terms of motion or dialogue. We argue a third question is missing: should the robot take its first hard-to-undo social action at all, such as a greeting, an uninvited grasp, or stepping into someone's space? We call this initiation authorization. Current frameworks rarely treat it as a separate safety layer. Today's stacks often skip this step: a high engagement score or a confident VLA rollout is treated as permission to act. But seeing a person is not the same as having their consent to be addressed. We frame initiation authorization within generalist-robot safety and contrast it with post-plan VLA guardrails, implementing PAS (probe-authorize-speak) on a doorway humanoid, comparing it with direct-init on logged traces, and proposing a three-condition user study, with open questions on metrics, governance, and where initiation ends and foundation-model generation begins.","short_abstract":"Safety for generalist robots is usually discussed in terms of motion or dialogue. We argue a third question is missing: should the robot take its first hard-to-undo social action at all, such as a greeting, an uninvited grasp, or stepping into someone's space? We call this initiation authorization. Current frameworks r...","url_abs":"https://arxiv.org/abs/2607.07420","url_pdf":"https://arxiv.org/pdf/2607.07420v1","authors":"[\"Zhijin Meng\",\"Francisco Cruz\"]","published":"2026-07-08T13:49:06Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"cs.HC\"]","methods":"[]","has_code":false}