13 Humanoid Deployments, Zero Complete Public Safety Files

13 current records, 0 complete public safety files under Biped's five-field disclosure test.

The disclosure gap is broad even where deployment evidence is comparatively strong. Biped could not fully verify any of the five checklist fields across the entire 13-record cohort. This is a transparency finding, not a finding that the robots are unsafe: operators and vendors may hold non-public risk assessments, workcell controls, training procedures, or incident processes.

The cohort is exactly the 13 current core deployment/test records from Biped Data Brief Issue 2. A record counts as a complete public safety file only when public evidence linked in Biped's canonical safety and deployment records fully discloses all five checklist fields. Partial descriptions and inferred operating practices do not count as full disclosure.

What zero means—and what it does not

Zero means the complete five-field package was not publicly verifiable. It does not mean the robots have no safety controls or that the deployments are unsafe.

The 13-record public disclosure matrix

Number of the 13 current core deployment/test records for which Biped could fully verify each public safety disclosure as of August 14, 2026.

Methodology

1. 1. Reuse the exact 13-record current core deployment/test cohort from Biped Data Brief Issue 2 so the safety analysis does not cherry-pick a different sample.
2. 2. For each deployment record, map the robot to Biped's canonical robot safety evidence and review the linked deployment source for five buyer-critical disclosures: emergency stop/stop authority, fall or power-loss behavior, human supervision/teleoperation, a named safety-standard or certification mapping, and incident/near-miss/corrective-action reporting.
3. 3. Count a field only when the public evidence fully discloses the relevant artifact or operating rule. A product specification, a named customer, a broad autonomy label, or an inferred workcell practice does not count as a full safety disclosure.
4. 4. Call a public safety file complete only when all five fields are fully disclosed for the same deployment record. None of the 13 records meets that threshold in the August 14, 2026 review.
5. 5. Keep deployment maturity separate from safety transparency: the GXO/Digit record remains production-status evidence even though its public safety file is incomplete under this disclosure test.

Limitations

• This is a public-transparency analysis, not a safety certification, hazard assessment, or conclusion that any robot is unsafe.
• Vendors, customers, insurers, integrators, and regulators may possess private risk assessments, work instructions, stop procedures, incident records, or conformity documentation that Biped cannot observe.
• The five-field checklist is an editorial buyer-diligence framework chosen for comparability; it is not an ISO, ANSI/RIA, OSHA, UL, NIST, or regulatory certification scheme.
• The cohort represents Biped's current tracked deployment/test records, not every humanoid deployment worldwide. Public evidence can change quickly, so each row carries both its deployment-source verification date and this safety-review date.

Frequently asked questions