From 16 to 18 October 2025 the 2025 World Intelligent Connected Vehicle Conference was held. The policy signals centred on three points: faster development of standards for combined driving assistance and autonomous driving, stronger production consistency supervision for vehicles with such systems, and tighter rules for over-the-air (OTA) software updates.
Why the focus shifts to responsibility
Intelligent connected vehicles are the most physical setting for AI: model output directly drives vehicle behaviour, and errors can involve personal safety. The compliance focus therefore naturally shifts from whether features are advanced to whether responsibility can be attributed.
The three signals map to three links in responsibility:
Standard-setting maps to design responsibility — what system design is safe and acceptable.
Production consistency supervision maps to manufacturing responsibility — whether production vehicles match the type-approved sample, avoiding hidden software or hardware configuration differences.
OTA rules map to update responsibility — a vehicle’s function can change after sale through remote update, which itself changes the risk state and must be traceable and controllable.
Why OTA is hard to regulate
OTA makes a vehicle a system that can keep changing, creating three supervisory difficulties.
Dynamic compliance status. The compliance state at launch may change with later updates, so supervisors need new mechanisms for continuous confirmation.
Timing of responsibility. When an accident occurs, which software version the vehicle was running and whether that version was adequately validated directly affect attribution.
Boundaries of consent. Whether updates touching driving functions require explicit user consent, and whether users can refuse, is where product design meets legal compliance.
What it means for companies
For vehicle and component makers: maintain a full-lifecycle ledger covering software versions, update records and validation reports. See our note on automotive standards for autonomous driving.
For AI system vendors: specify model updates, data feedback and accident liability allocation in contracts with OEMs. See our note on AI commercial contract clauses.
For data and security teams: road data and in-vehicle data involve personal information and important data, and their processing and cross-border transfer must comply with the relevant rules. See our note on network data security risk assessment and our data compliance service.
AI’s convergence with automotive brings software compliance complexity into manufacturing. Companies need to convert the convenience of updatable software into the capability of traceable states. See our note on the AI plus transportation action plan.
This article is based on public information from the 2025 World Intelligent Connected Vehicle Conference. It is for general reference only and does not constitute legal advice.