New reports are putting Tesla’s driver-assistance technology back under the microscope as regulators examine how drivers use it on American roads.

A car can keep moving while its driver appears to be asleep.

That startling possibility is at the center of a new Tesla FSD safety controversy this week. A report cited by Reuters identifies 43 videos showing Tesla drivers apparently asleep or disengaged while their vehicles were moving, including 17 videos from 2026. The findings have prompted Rep. Raja Krishnamoorthi of Illinois to urge the U.S. Transportation Department to take action.

 

The videos don’t prove that Tesla’s software caused crashes or that every driver in them was actually asleep. But they expose a much bigger problem: people may be treating driver assistance as if it were a chauffeur.

Tesla calls its system Full Self-Driving, but the name can create a dangerous misunderstanding.

Tesla’s FSD Supervised system still requires the driver to remain attentive and ready to take control. Federal safety documents also describe FSD as an advanced driver-assistance system, rather than a fully autonomous system.

Think of it like cruise control with a very sophisticated brain. You can let the technology handle more of the driving, but you haven’t handed over responsibility for the trip.

That’s why the latest reports matter beyond Tesla’s fan base. If drivers believe the car is capable of handling everything, they may stop behaving like drivers long before the technology is actually capable of handling every situation.

And regulators are already looking closely. NHTSA has been investigating Tesla’s FSD technology, including questions about how the system performs when visibility deteriorates. Its investigators have raised concerns about situations involving glare and other conditions that can interfere with camera performance.

For an ordinary American driver, this isn’t really about Silicon Valley or artificial intelligence.

It’s about the person commuting home after a long shift, the parent carrying kids across town, or the family taking a six-hour highway trip.

Most drivers already understand that falling asleep behind the wheel is dangerous. NHTSA says drowsy driving contributed to 644 deaths in 2024, and the agency warns that fatigue can damage attention, judgment and reaction time.

But advanced driver assistance creates a strange psychological trap. The smoother the car feels, the easier it can become to forget that you’re still responsible for what happens next.

That’s the part that should make drivers stop and think.

Imagine riding in an elevator that suddenly started steering itself through the building. You might trust the machinery, but you’d still want to know exactly what happens when something goes wrong.

A highway is far less forgiving.

For families, the stakes are even higher. A driver who misunderstands the limits of FSD isn’t only putting one person at risk. Passengers, pedestrians and other motorists are sharing the same road.

And there’s a financial consequence, too. A serious crash can mean medical bills, lost work, higher insurance costs and a damaged vehicle. Technology that promises convenience can become very expensive when a driver assumes it can do more than it actually can.

Critics argue that Tesla has a responsibility to make sure drivers don’t become overconfident.

That argument isn’t coming from nowhere. NHTSA has previously raised concerns about driver engagement and Tesla’s public messaging around FSD, including whether some messaging could encourage people to view the system more like a chauffeur than a driver-assistance tool.

At the same time, there’s another side to the debate.

Advanced driver-assistance technology can potentially reduce human mistakes, and NHTSA’s newer testing program is explicitly designed to evaluate these systems and give consumers more information about their safety performance. A 2026 Tesla Model Y became the first vehicle to pass the agency’s new ADAS benchmark.

That creates the central tension.

The technology may become increasingly capable while the human driver becomes increasingly tempted to stop paying attention.

And that’s a problem no software update can solve by itself.

The next phase of this story will likely focus on what Tesla’s safeguards can actually detect and how effectively they prevent misuse.

Rep. Krishnamoorthi has asked federal officials to examine whether Tesla’s safeguards can be defeated or bypassed, including questions involving cabin-camera monitoring.

Meanwhile, Tesla’s broader push toward autonomous vehicles is making the debate even more important.

NHTSA opened an audit in September into Tesla’s self-certification of the Cybercab after the driverless vehicle was deployed in Austin, Texas. That investigation shows how quickly the federal government is confronting a transportation world that looks very different from the one existing vehicle regulations were designed around.

That’s the fascinating part.

We’re watching two technologies develop at the same time: cars that can handle more of the driving and humans who may be willing to hand over more control.

The question isn’t simply whether Tesla can make its software smarter. The harder question is whether the technology can keep humans appropriately engaged while it does.

The future of driving may depend less on how smart the car becomes than on whether people remember they’re still behind the whee