European approval of Tesla’s Full Self-Driving Supervised technology is placing commercial confidentiality and regulatory transparency in tension as authorities consider whether the driver-assistance system should be permitted across the European Union.
The Netherlands Vehicle Authority, RDW, has already granted a European type approval that currently allows the system to be used in the Netherlands. Wider use across the EU would require further approval through a European process involving the European Commission and member states.
Scrutiny has increasingly focused on the safety evidence behind that decision. European regulators have declined to publish detailed testing information and methodologies where they consider the material commercially sensitive, including information supplied by Tesla during the approval process.
RDW says manufacturers routinely provide confidential technical material during vehicle approvals and that the legal framework can prevent it from making all substantive evidence public. Following scrutiny in the Dutch parliament, however, the regulator has released more information about the work behind Tesla’s approval.
The authority says it assessed the system for more than 18 months, conducting more than 3,000 hours of testing and more than 1,000 test runs across test tracks and public roads. Its assessment included data from 1.8 million kilometres driven in Europe with Full Self-Driving Supervised.
RDW also says it independently reviewed Tesla’s statistical analysis, verified how the manufacturer collected its data and assessed whether the underlying information was complete, verifiable and accurate.
“Safety is the RDW’s foremost priority,” the authority said. It has described the European version of the system as “at least as safe as other driver assistance systems” when operated under the conditions of its approval.
Full Self-Driving Supervised remains a driver-assistance product rather than an autonomous-driving system. The driver remains responsible for the vehicle and must be ready to intervene, while sensors monitor attention and the system can issue warnings or prevent continued activation when the driver repeatedly fails to respond appropriately.
RDW says almost 40,000 Teslas fitted with the system had collectively travelled about 24 million kilometres in the Netherlands by mid-June without a relevant incident. It has also increased the reporting frequency required during post-approval monitoring from annually to monthly.
Before the system can be accepted more widely, RDW must submit an application to the European Commission and the proposal must pass through the relevant member-state process. Approval would significantly enlarge the market available for Tesla’s software-based driving functions.
The regulatory questions extend beyond one manufacturer. As software and artificial intelligence take a larger role in regulated products, traditional approval processes are having to accommodate systems whose performance depends heavily on data, statistical comparison and algorithms that continue to evolve after a product reaches customers.
Vehicle manufacturers have long supplied regulators with proprietary engineering information, much of which receives legitimate protection as commercially sensitive intellectual property. Software-led products place greater weight on testing methodology and data analysis, however, creating stronger pressure for outside scrutiny of the evidence regulators use.
The same tension is emerging in medical technology, financial services and industrial automation, where proprietary models can increasingly influence regulated outcomes. Authorities must protect trade secrets while maintaining enough transparency to demonstrate that safety and compliance decisions rest on credible evidence.
Automotive regulation presents a further complication because software can continue to change after type approval. Over-the-air updates allow manufacturers to alter functionality without the traditional product-cycle boundaries associated with mechanical systems, increasing the importance of post-market monitoring alongside initial certification.
Competition is also intensifying. Tesla is trying to protect its European position as established manufacturers and newer Chinese entrants expand their electric-vehicle ranges, while advanced driver-assistance software has become another area in which carmakers can differentiate products and create recurring revenue.
BMW and Ford already hold approvals for advanced assistance functions under different conditions, meaning decisions taken around Tesla will sit within a broader European framework rather than create a completely new category of regulation.
Automotive groups are likely to watch closely how much evidence regulators are ultimately prepared to disclose, particularly where approval depends on proprietary datasets that competitors could seek to examine. Regulators, meanwhile, face a growing expectation that decisions involving automated systems should be understandable outside the manufacturer-authority relationship.
As vehicles become more software-defined, confidence in the approval process will depend not only on the technical performance of individual systems, but also on whether regulators can explain how that performance has been tested without disclosing intellectual property they are legally required to protect.
You must be logged in to post a comment.