CLEPA response to the public consultation on Artificial intelligence
On 19 February, the European Commission adopted a White Paper on Artificial Intelligence, which sets out options for a regulatory framework, expected towards the beginning of 2021. The Commission considers imposing mandatory requirements for AI applications that would be considered “high-risk,” and proposes to review existing legislation on consumer protection, product safety and liability, data protection, and privacy, to better take into account AI-specific concerns.
CLEPA submitted a response to the public consultation on the White Paper, outlining the automotive suppliers’ position on AI and on the proposed regulatory changes.
Summary of the CLEPA response
CLEPA welcomes the White Paper and supports the risk-based approach proposed by the Commission, with mandatory ex-ante requirements for high-risk applications. In this regard, while we agree that transport should generally be considered high-risk, we nevertheless stress that some transport-related applications may not be high-risk and should not be subjected to the mandatory ex-ante requirements.
We underline the importance of not hindering innovation unnecessarily. Requirements should always remain technically achievable and proportionate to the possible risks, leaving room for testing and experimenting.
With regards to the upcoming legislative framework, CLEPA believes that a sector-based approach would be the most appropriate. The automotive sector is very specific in that it is already subject to strict ex-ante conformity controls, through the type approval framework. Vehicles must undergo a strict testing and certification process before they can be placed on the EU market, to ensure that their safety is compatible with the latest standards. The new AI-related requirements should be included into the existing type approval framework, rather than as a new set of requirements controlled separately. It is essential that certification, testing, and market surveillance are not duplicated.
The EU legislative framework on AI and the UNECE requirements for Automated Driving Systems (ADS) should be aligned, with future UNECE requirements to be considered valid AI-related requirements, rather than adding another regulatory layer.
We believe that the existing EU legislative framework on security, liability, and responsibility is already largely fit-for-purpose and should not be completely redrawn for AI. Nonetheless, it should be carefully assessed whether and to what extent AI applications and their specificities are already addressed by the current liability framework.
