Tesla’s Full Self-Driving (FSD) system was tested daily on Queensland roads over 100 days, recording more than 500 safety-related interventions. While the technology shows advanced capabilities, significant limitations remain—many tied to road design that today favors human drivers over machines.
- Over 500 safety-critical interventions recorded in 100 days of Tesla FSD testing.
- Road elements like school zones, railway crossings, and zipper merges cause repeated system confusion.
- Improved road markings, clearer signs, and consistent infrastructure required for safer autonomy.
What happened
During a 100-day testing period on Queensland roads, Tesla’s Full Self-Driving system was put through more than 500 safety-critical events where human driver interventions were necessary. The Tesla Model Y handled some situations with a high degree of precision, even surpassing typical human driving performance in certain tasks. However, numerous challenges surfaced, particularly in interpreting complex or ambiguous road environments, such as school zones with time-dependent speed limits and peculiar local traffic rules like zipper merges.
The system also exhibited notable vulnerabilities at infrastructure points like small bridges, railway crossings, and challenging roundabouts common in Australia. It misidentified certain road users, including e-scooter riders, and struggled under adverse weather conditions. These findings highlighted recurrent issues revealing that current road designs present significant barriers for fully autonomous operation.
Why it matters
The test shows that autonomous driving technology in Australia is not yet mature enough to operate safely and independently across the variety of real-world road scenarios it encounters. Although vehicles like Tesla’s FSD demonstrate promising capabilities, their limitations in handling common but subtle driving conditions raise safety concerns. The need for frequent driver intervention remains a barrier to full autonomy and public acceptance.
Furthermore, the findings emphasize that technology alone cannot address these challenges. Australian roads, bridges, and intersections were designed for human cognition and behavior, not for artificial intelligence. For robotaxis and autonomous vehicles to operate efficiently and safely, infrastructure must evolve in tandem—featuring clearer markings, consistent signage, and less ambiguous designs that accommodate machine interpretation.
What to watch next
Future progress in Australian autonomous driving will hinge on collaborative efforts between vehicle technology developers and infrastructure planners. Authorities might prioritize basic yet impactful upgrades such as more frequent and consistent speed limit signs, enhanced lane markings, and road layouts that reduce ambiguity. These moderate road improvements could significantly improve AI system reliability without heavy investment in costly smart-road installations.
Meanwhile, ongoing public trials and archives like the White Box Autonomy initiative will continue to gather essential data on autonomous vehicle performance in diverse conditions. This feedback loop will be crucial in defining regulatory frameworks and safety standards before large-scale deployment of robotaxis can proceed. Monitoring how other autonomous vehicle operators, such as Waymo, address these challenges in Australia will also be key.