Proposal to delete a paragraph that should have been deleted when the informal group approved the proposals in document ADS-14-38.

Proposal to amend ADS-20-12 (consolidated guidance document). The terms faults, malfunctions, and abnormalities lack distinct meanings in requirement 6.1.4.2. The revised explanation clarifies that “faults, malfunctions, and abnormalities” do not have specific distinct meanings and cover any failure in the ADS’ capability to perform the DDT within the ODD. The three kinds of issues encompassed are warning signals from other systems, faults without warning signals a driver would perceive, and faults within the ADS itself.

Proposal to amend ADS-20-12 (consolidated guidance document). The ADS shall have strategies to detect and respond to instructions from road safety agents. For ADSF-1, this allows transfer of control to the fallback user. Manufacturers may use non-ADS strategies including remote intervention or permitting the road safety agent to drive. Examples of instructions include pull over, give way, stop after collision, move out of way, avoid specific road, and not overtake. Instructions may be issued via lights, sirens, audio messages, written signs, or light bar messages.

Proposal to amend Paragraph 5.3.1.3 (GTR) / 7.3.1.3 (UNR) in Safety Case. The revision replaces “no new behaviours or traffic flows” with “no new behaviours or traffic flows density or dynamic element types” and adds “or traffic behaviour” to the text regarding traffic laws because ODD does not include a definition of traffic behaviour, making the original sentence factually incorrect.

Proposal to amend para. 6.2.2.3.8.1 (b) to change “a driving task relevant area” to “the driving task relevant area” and to expand the accompanying GID text to define the driving task relevant area as consisting of different zones including inside mirror, left outside mirror, right outside mirror, left window, right window, windscreen, and instrument cluster or instrument display, and to clarify that depending on the type of ADS feature and driving situation the user could be suitable to have looked at just one zone or more than one zone.

Proposal to amend explanation of para. 6.2.2.1.4. to clarify that ‘while’ refers to continuous presentation of information sufficient to enable a user to take over performance of the DDT such as through ambient lighting, tell-tales, or dashboard colour changes. The explanation notes that information on adapted performance is only presented if applicable.

Proposal to revise the guidance on para. 6.1.2.3 of the draft UN Regulation on ADS regarding unreasonable disruption to traffic. The revision changes “prevent normal interaction” to “allow normal interaction” and “require other road users” to “may technically require other road users” to clarify that the requirement permits normal interaction with traffic, such as using on ramps, slowing for traffic lights, and waiting at unprotected turns, where other road users must adapt their speed, without violating the requirement.

Proposal for guidance on operational states, including remote termination and a diagram illustrating ADS and feature operational states, update guidance on terms “on”, “off”, “available,” “active,” “unavailable,” and “disabled” as canonical operational state terms for ADS descriptions and clarify that “active” applies only to features.

Proposal to replace text explaining that the ADS IWG did not reach consensus on definitions with clarification that remote intervention and remote interaction were not intended to convey different meanings in the regulation, both covering cases where the ADS requires support for tactical functions while continuing to perform the entire dynamic driving task, note that this aspect of ADS is a developing area, and include references to SAE J3016-2021, EU 2022/1426, the Automated Vehicle Safety Consortium Best Practice for ADS Remote Assistance Use Case (Nov. 27, 2023), and the upcoming ISO/PAS 26361 standard.

Proposal to add a new section on Cross-Cutting Organisational and Assurance Risks to AI-07-06. Six risks are proposed: insufficient organisational governance for AI safety management; inadequate definition of roles, responsibilities, and competencies; lack of independence between development and verification activities; degradation of AI safety processes over time; ineffective change management for AI-related artefacts and processes; and inadequate management of AI supply-chain risks. Each risk addresses factors across the lifecycle that can influence the effectiveness of technical AI risk mitigations.