WP.29-199-17
AI informal group status report to WP.29

The AI IWG held two meetings since the last WP.29 session and has virtual meetings scheduled for 23 July 2026 and 3 September 2026. The IWG compiled literature on regulated automotive safety systems using AI (AI-09-04), created Terms & Definitions for AI (AI-09-05), listed AI use cases for regulated automotive safety systems (AI-09-03), and created a draft guiding questions document (AI-08-07). A working draft AI Reference Document (AI-09-06/Rev.3) exists but is not finished. The AI IWG requests an extension of its mandate to deliver the AI Reference Document as a working document no later than June 2027.

UNECE server
Excerpts from session reports
WP.29 | Session 199 | 22-26 Jun

Informal working group on Artificial Intelligence
21. The representative of the United Kingdom of Great Britain and Northern Ireland introduced the status report of the Informal Working Group on Artificial Intelligence (WP.29-199-17), highlighting the activities carried out to date under paragraphs 6(a) to 6© of the terms of reference. He presented the draft Reference Document for the Use of Artificial Intelligence (AI) in Regulated Automotive Safety Systems (WP.29-199-26) and proposed to extend the mandate of the group by one year or more in order to allow the submission of a working document. He also drew attention to possible areas of work not yet included in the current terms of reference. Several delegations expressed support for extending the mandate by one year to enable completion of the task defined under paragraph 6(d) of the terms of reference. The representative of Canada suggested that the mandate could be extended by two years.

22. WP.29 agreed to extend the mandate of the Informal Working Group on Artificial Intelligence by one year to complete its current tasks and invited the group to consider next steps, as well as potential new areas of work for consideration by WP.29 at its June 2027 session. The representative of the Netherlands (Kingdom of the) expressed his concern on the use of AI in relation to cyberthreats and asked for timely attention on this thread within the IWG on Artificial Intelligence