Tesla Claims FSD Supervised Avoided Deer Collision in Direct Sunlight, But Evidence Remains Mixed

There’s a moment every rural driver dreads. You’re moving at speed, the sun is low and blinding, and something darts into the road. For most of us, it’s pure instinct and luck. Tesla is now suggesting its Full Self-Driving (Supervised) software can handle exactly that situation — but the full picture is rather more complicated.

Tesla’s official X (formerly Twitter) account posted a clip showing one of its vehicles using FSD Supervised to detect and brake for a deer crossing the road, with the caption stating that “FSD Supervised saved a deer because it was able to see through direct sunlight.” The post has drawn attention across the automotive and technology press, reigniting debate about how reliable camera-only driver assistance really is when conditions get difficult.

What Is FSD Supervised, and How Does It Work?

It’s easy to get lost in Tesla’s branding here, so a quick explanation is useful. FSD Supervised — Full Self-Driving (Supervised) — is what’s known as a Level 2 driver-assistance system. That means it can handle steering, acceleration and braking in many situations, but it is not autonomous. The driver must remain fully attentive, keep their hands on the wheel, and be ready to take control at any moment. Tesla is clear about this in its own documentation. The name might sound like the car is driving itself, but legally and practically, the human behind the wheel is still responsible.

What makes Tesla’s approach distinctive — and controversial — is that newer vehicles rely entirely on cameras. From around 2022 to 2023, Tesla removed radar from its new cars and moved to what it calls “Tesla Vision,” a camera-only system backed by neural networks. The company argues that vision-based AI, trained on vast amounts of real-world driving data, can outperform traditional sensor combinations. Critics disagree, pointing to radar and LiDAR as tools that work in conditions where cameras struggle — fog, heavy rain, darkness, and yes, direct sunlight.

The Case For and Against

The tweet fits into a broader pattern. Tesla and its supporters regularly share user-submitted clips showing FSD successfully avoiding hazards — pedestrians stepping off pavements, vehicles pulling out unexpectedly, and now a deer in glare conditions. And there are genuine examples of the system detecting animals and braking before impact, including on motorways and in low-light scenarios.

But it’s not the whole story.

Separate documented incidents — captured on dashcams and reported across automotive and fact-checking outlets — show Tesla vehicles on FSD or Autopilot striking deer without braking or slowing at all. Drivers in those cases reported that the system simply didn’t register the animal in the road. The US National Highway Traffic Safety Administration (NHTSA) has ongoing interest in Tesla’s driver-assistance systems and has previously opened investigations into Autopilot and FSD crashes, with object recognition and reliability forming part of that wider scrutiny.

Dan O’Dowd, founder of the Dawn Project, a campaign group that has repeatedly tested and criticised Tesla’s FSD software, said: “One video of a success does not tell you the failure rate.”

That tension — between compelling individual clips and the absence of systematic, independently verified testing data — sits at the heart of this debate. Tesla processes multi-camera feeds at high frame rates and uses neural networks to identify objects, but no independently verified accuracy figures for wildlife detection have been published by UK or EU regulators.

The Sunlight Problem

Direct sunlight is a genuine technical challenge for camera-based systems. When a camera is hit by strong glare or a low sun angle, the sensor can be saturated — essentially blinded — in the same way a human driver might be. High contrast between a bright sky and a dark road surface also makes it harder to distinguish objects. Tesla’s tweet is essentially claiming that its AI can handle this edge case reliably. But that claim rests on a single user video, not controlled testing. The exact software version, hardware generation, and location shown in the clip haven’t been independently verified.

Competing manufacturers largely use sensor fusion — combining cameras with radar and sometimes LiDAR — precisely to cover these kinds of gaps. When a camera is overwhelmed by glare, radar can still detect a solid object in the road. Tesla’s bet is that sufficiently advanced vision AI makes that redundancy unnecessary. Whether that bet pays off consistently, across all conditions and all drivers, remains an open question.

Wildlife on UK Roads

Animal collisions aren’t a minor issue. Department for Transport figures put police-reported road traffic accidents involving animals at roughly 1,000 to 2,000 per year in Great Britain, with a much larger number of minor, unreported incidents. Deer are among the most common large animals involved, and collisions can cause serious vehicle damage and injuries to drivers and passengers. Driver-assistance systems with automatic emergency braking and object detection are increasingly promoted by road safety organisations as tools that could reduce these incidents — but only if they work reliably.

What This Means for Kent Residents

Kent’s mix of rural B-roads, faster A-roads like the A21 and A249, and areas of countryside where deer are regularly spotted means wildlife collisions are a real, local risk — and any driver using FSD Supervised here must remember that UK law places full responsibility on them, not the software. The mixed evidence about Tesla’s deer detection means it cannot be treated as a guarantee, chiefly in glare or poor weather. If you’re driving a Tesla in Kent with FSD Supervised active, staying alert and ready to take over isn’t just good advice — it’s a legal requirement.