Tesla's FSD System Under Scrutiny in Brussels for Speeding Ahead of EU Vote
A recent investigation in Brussels has brought Tesla's "Full Self-Driving (Supervised)" system under intense scrutiny, revealing a concerning propensity for exceeding speed limits in designated low-speed zones. This study, highlighting instances where the system significantly surpassed the 30 km/h limit in over half of the tested areas, has ignited a debate about autonomous driving safety. The timing of these findings is particularly critical, as European Union member states are poised to cast their votes on whether to grant FSD widespread approval across the entire bloc, with several nations already expressing reservations about its speed control capabilities.
The “FSD30” study, spearheaded by Johanna.be, a non-profit organization based in Brussels dedicated to enhancing safety for pedestrians and cyclists, documented numerous instances of the FSD system failing to adhere to local speed regulations. Over three days in July, researchers traversed approximately 400 kilometers, focusing primarily on urban streets where vulnerable road users are most prevalent. A key issue identified was the system's inability to correctly interpret speed limits; often displaying a 50 km/h limit on the vehicle's screen when the actual regulatory speed was 30 km/h. This discrepancy resulted in an average speed of 44 km/h in these restricted zones, a rate that significantly elevates the risk of fatal injury to pedestrians, according to data from the World Health Organization.
Beyond the core issue of speeding in 30 km/h zones, the study also pinpointed problems within “woonerven,” or residential shared streets, where FSD vehicles exceeded the 20 km/h limit in 42% of cases. Additionally, the system was observed attempting to overtake cyclists in “fietsstraten” (cycle streets), a practice explicitly forbidden under Belgian traffic laws. While researchers noted FSD's commendable tendency to yield to pedestrians and cyclists at intersections, they concluded that such courtesies do not mitigate the systemic issue of excessive speed in environments designed for heightened safety. These findings have been submitted to relevant Dutch and Flemish authorities, urging them to re-evaluate their prior endorsements of the system ahead of the crucial EU vote.
Despite previous provisional approvals by certain countries, including Belgium, which greenlit FSD (Supervised) after 5,000 km of testing on Flemish roads, the recent study underscores the system's persistent challenges. A spokesperson for the Flemish government, when questioned about the study, reiterated that the system is still in a 'learning' phase and that drivers bear '100% responsibility.' However, critics argue that this stance shifts accountability from the software's design flaws to the end-user, posing a significant ethical and safety dilemma. The upcoming EU vote, scheduled to take place within the Technical Committee on Motor Vehicles (TCMV), will require a qualified majority for FSD to receive bloc-wide recognition, with speeding emerging as the primary contention point for several member states, including Sweden and France.
Tesla has attempted to counter these criticisms with its own data, asserting that FSD vehicles do not speed excessively and generally match the flow of surrounding traffic. This argument, however, appears to miss the mark in contexts like Brussels, where speed limits were intentionally reduced to address human drivers' historical tendency to drive too fast. Furthermore, past instances have revealed Tesla's reliance on self-published safety data that some have deemed misleading, potentially inflating FSD's safety performance through skewed comparisons. The ongoing debate highlights the complex challenges in integrating advanced autonomous driving technologies into diverse urban environments and regulatory frameworks.
The core message emerging from this research is that Tesla’s autonomous driving system struggles with a fundamental task: accurately identifying and adhering to posted speed limits. This deficiency is not merely an occasional glitch but a systemic flaw, particularly problematic in dense urban areas populated by pedestrians and cyclists. The notion that drivers are solely responsible for rectifying the software's errors when the system itself misinterprets crucial traffic regulations represents a significant challenge to the widespread acceptance and deployment of autonomous vehicles.