UL 4600 is the one published standard that asks an autonomous product for a safety case

The document exists. The ledger fetched it at its publisher and quotes it.

Record ID

PAG-2026-0025

Kind

Standard

Jurisdiction

Global

Last verified

15 September 2026

Added

15 September 2026

• The UL Standards catalogue lists UL 4600, Standard for Evaluation of Autonomous Products, as active, edition 3, with a published date of 17 March 2023 and ANSI approval the same day.
• UL states the standard calls for manufacturers to consider safety case scenarios so that autonomous trucks and cars can safely share the road.
• Earlier editions 1, 2 and 2-1 are listed in the same catalogue, with edition 2-1 also marked active.

The machine, The AI inside it, Liability. Stated means the document you can open below says it; silent means the ledger read the document and it does not.

The machineStated

The unit of compliance is an argued safety case with evidence, not a checklist of design features.UL Standards and Engagement, UL 4600 Standard for Evaluation of Autonomous Products, primary source, 17 March 2023.

The AI inside itStated

Machine learning components are addressed as claims to be supported rather than as a component with a failure rate.UL Standards and Engagement, UL 4600 Standard for Evaluation of Autonomous Products, primary source, 17 March 2023.

LiabilitySilent

A safety case is a document a court can read, which is why insurers ask for it, though the standard imposes no liability rule.UL Standards and Engagement, UL 4600 Standard for Evaluation of Autonomous Products, primary source, 17 March 2023.

If nobody supervises your machine, you will be asked how you know it is safe, and this is the published answer format. Building the safety case as you develop is cheaper than reconstructing it for a customer, an insurer or an exemption petition, and it is the only such framework with an ANSI approved edition on the shelf.

1. UL Standards and Engagement, UL 4600 Standard for Evaluation of Autonomous ProductsPrimary · 17 March 2023

• VerifiedStandardPAG-2026-0024

  The standard for harm with no broken part is SOTIF, written for road vehicles and borrowed by everyone else

  Global · verified 15 September 2026

  ISO 21448:2022 is edition 1, published 2022-06, 181 pages, from ISO technical committee 22, and ISO marks it at stage 90.92, to be revised.

• Open questionOpen questionPAG-2026-0032

  Open question: how do you evidence a safety case for a controller that was learned rather than specified?

  Global · verified 15 September 2026

  ISO 13849-1 and IEC 62061 assume a specified safety function with quantifiable reliability in high or continuous demand mode.

• Open questionOpen questionPAG-2026-0029

  Open question: who is liable when a behaviour the robot learned causes the harm?

  Global · verified 15 September 2026

  The AI Liability Directive proposal, which would have eased proof for exactly this case, was withdrawn and published as withdrawn in the Official Journal on 6 October 2025.

• VerifiedStandardPAG-2026-0028

  No published safety standard covers a legged humanoid, and the one that would is still a committee draft

  Global · verified 15 September 2026

  The ISO TC 299 catalogue lists ISO/CD 25785-1, Robotics, Safety requirements for dynamically stable industrial mobile robots, Part 1: Robots, at stage 30.60.

• VerifiedStandardPAG-2026-0027

  The AI management system standard is the one certificate a robot maker can actually hold today

  Global · verified 15 September 2026

  ISO/IEC 42001:2023 is edition 1, published 2023-12, 51 pages, at stage 60.60, from ISO and IEC joint technical committee 1 subcommittee 42.

• VerifiedStandardPAG-2026-0026

  IEEE 7007 is an ontology for ethically driven robotics, and it is the only active standard of its kind

  Global · verified 15 September 2026

  IEEE lists IEEE 7007-2021 as an active standard: a set of ontologies with different abstraction levels containing concepts, definitions, axioms and use cases for ethically driven design.