As Computers Take the Driver’s Seat, Experts Worry About Safety, tell me about this news in (https://www.nytimes.com/2026/09/22/business/self-driving-cars-tesla-fsd.html)

The report titled “As Computers Take the Driver’s Seat, Experts Worry About Safety” examines the escalating safety debates and regulatory scrutiny surrounding autonomous vehicle technology, with a particular focus on Tesla’s Full Self-Driving (FSD) system and the broader push toward fully driverless robotaxis.

The key points and themes highlighted in the coverage include:

1. The Gap Between Promises and Real-World Reliability

While automakers and autonomous technology advocates argue that computer systems will dramatically cut traffic fatalities by eliminating human errors like distraction, drowsiness, and intoxication, safety researchers and transportation experts caution that automated systems are not yet sufficiently reliable. Edge casessuch as complex construction zones, emergency vehicles, erratic pedestrian behavior, and unusual road debriscontinue to trigger unpredictable software behaviors and sudden system disengagements.

2. Tesla’s Vision-Only Strategy Under Scrutiny

A central focus of expert concern is Tesla’s unique reliance on a camera-only (vision-based) approach paired with end-to-end neural networks, which foregoes hardware like LiDAR and radar as well as pre-mapped geofences. While competing autonomous fleets (such as Waymo) use multi-sensor redundancy and strict geofencing to maintain safety margins, experts warn that camera-only systems remain vulnerable to sensor blinding from intense glare, severe weather (heavy rain, fog, snow), and optical illusions.

3. The Automation Complacency Dilemma

For systems marketed under names like “Full Self-Driving” that remain Level 2 driver-assist platforms requiring constant human attention, safety advocates point out a persistent psychological hazard: driver complacency. Because the software handles driving duties smoothly for long stretches, drivers tend to disengage mentally, check their phones, or fail to intervene in the fractions of a second needed when the system makes an abrupt error.

4. Contested Safety Metrics and Crash Data

The article highlights growing pushback against how automakers present autonomous safety data. Critics and independent analysts note that comparing crash statistics of newer, advanced vehicles primarily operating on highways against the entire aging U.S. vehicle fleet produces skewed results. Federal regulators, including the National Highway Traffic Safety Administration (NHTSA), have faced calls to mandate more rigorous, standardized reporting for autonomous incidents and near-misses.

5. Regulatory Challenges and Driverless Deployments

As companies push to eliminate physical controlssuch as pedals and steering wheelsin upcoming purpose-built robotaxi designs, regulators and municipal officials are caught between encouraging technological innovation and safeguarding the public. Experts stress that without uniform federal benchmarks and transparent third-party safety validation, putting automated software fully in charge of passenger safety poses significant public risk.