Tesla's most recent software update for its Full Self-Driving system, identified as version 14.3.9, has reportedly caused an unexpected issue for vehicle owners. The update is triggering erroneous alerts that indicate dirty cameras, despite the cameras being clean and fully functional. This glitch has led to a number of drivers scheduling unnecessary service appointments, which the automaker has subsequently started canceling as it acknowledges the software defect. A solution for this problem is anticipated in a forthcoming system update.
The problem surfaced following the release of FSD update v14.3.9, which encompasses software builds 2026.27.5 and 2026.27.6, rolled out around September 12. Owners who installed this version began receiving a persistent message on their central display: "Poor camera visibility detected. Interior windshield residue cleaning is required." Curiously, this alert often appears in conditions of bright sunlight or glare, suggesting the software misinterprets normal optical circumstances as an obstruction to the forward-facing camera system. Social media posts from Tesla owners, including one from @CARN0N on X (formerly Twitter), showcased Tesla Support confirming the software fault and the company's efforts to preemptively cancel service bookings related to this false alarm. Similar reports from other owners on platforms like Reddit further corroborate the widespread nature of this issue.
The FSD v14.3.9 update also introduced a significant enhancement: Automatic Collision Evasion for vehicles equipped with Hardware 4, enabling the FSD system to intervene and prevent collisions during manual driving. However, the camera alert bug has overshadowed this improvement. Tesla advises owners experiencing the false warnings to refrain from scheduling service appointments. For those who have already incurred costs for inspections or cleaning services due to these alerts, Tesla Support is offering refunds. A subsequent update, v14.3.10, is expected to resolve the erroneous notifications. While its release notes currently mirror those of v14.3.9, it is anticipated to contain the necessary corrections.
This incident underscores the inherent complexities and ongoing challenges in perfecting vision-only autonomous driving systems. While a false camera warning may seem a minor inconvenience rather than a safety hazard, it highlights the delicate balance required for such systems to accurately distinguish between genuine environmental degradation (like dirt, fog, or dust) and benign optical conditions. The system's heightened sensitivity, which appears to be the root cause of this particular bug, indicates an overshoot in calibration. Achieving reliable and unsupervised self-driving capabilities at scale necessitates a robust and redundant sensing architecture. The current reliance on a camera-only approach, despite its advancements, is prone to such misinterpretations, suggesting that continuous refinement and potentially diverse sensor inputs will be crucial for overcoming these hurdles on the path to widespread autonomous vehicle deployment.