Tesla’s “Full Self-Driving” (FSD) system recently hit headlines after a vehicle, reportedly operating on FSD mode, struck a deer without slowing down or stopping. Tesla’s unique approach to automotive autonomy relies solely on data from cameras, unlike other systems that employ sensors like LiDAR for added accuracy. Instead of utilizing these 3D-detection sensors, Tesla’s FSD relies on image recognition to identify and respond to obstacles. However, this recent incident highlights the risks associated with Tesla’s strategy.
Footage posted by @TheSeekerOf42 on X (formerly Twitter) shows the car failing to detect a deer standing in the roadway. The video captures the vehicle’s approach without any signs of slowing, followed by images of a dented hood and cracked bumper—damage suggesting a significant impact. Despite the collision, Tesla’s system reportedly did not initiate an emergency stop or relinquish control back to the driver, raising concerns about the car’s response protocols in unexpected situations.
The vehicle owner, who subscribes to Tesla’s FSD service, labeled the incident an “edge case.” However, many argue that true autonomy should account for such “edge cases” since these unpredictable scenarios are precisely where drivers rely most on the system’s performance. With its current technology, Tesla’s camera-based FSD struggled to recognize the light-colored deer against the road, a situation where LiDAR could have provided essential depth perception to recognize the animal.
What’s more troubling is the system’s lack of post-collision response. Typically, an impact sensor or camera should prompt the vehicle to slow down or pull over, especially when significant damage occurs. Tesla’s design choices, prioritizing cost and efficiency over additional sensory tech, have led to questions about whether it’s prepared to handle real-world complexities.