13. The expert from France shared the concerns of his delegation regarding transparency, the basis for building trust in AI technologies. He stressed the importance of being able to check the good functioning of such systems. He volunteered to elaborate on this statement at the next GRVA session.

14. The expert from OICA raised questions for clarification and asked why FIA preferred narrow AI and supervised learning (FIA answered that it was a conservative position). He agreed with FIA regarding online learning: he supported the view of GRVA that a new version of a type-approval relevant software based on AI was subject to approval, according to the provisions of UN Regulation No. 156, before deployment.

15. The expert from the Russian Federation supported that this matter was addressed by the IWG on Functional Requirements for Automated and Autonomous Vehicles (FRAV). The expert from OICA/CLEPA argued that this was also relevant for the IWG on VMAD.

16. The expert from Germany mentioned the importance of this work: sound definitions were the prerequisite before starting regulatory initiatives. He offered to comment the definitions proposed.

17. The expert from the United States of America questioned the need to regulate AI.

18. The expert from ITU stated the importance of understanding how AI works. He also stressed that software using AI would need to be frozen and tested before being deployed.

19. The expert from United Kingdom of Great Britain and Northern Ireland offered views on what could be subject to regulations. He mentioned as an example the risks related to discrimination. He suggested that future steps would be needed in the near future.

20. GRVA agreed to keep this item on its agenda. GRVA invited the expert from OICA to organize a workshop to finalize the work on definitions, with the assistance of the secretariat.

21. The expert from SAE International volunteered his members to assist in that process. GRVA invited SAE International to join the workshop.