Tesla’s ‘self-driving’ software fails at train crossings, car owners warn
Some Tesla owners are raising serious safety concerns about the company’s “Full Self-Driving” (FSD) software, claiming it fails to properly recognize and respond to railroad crossings — a basic but potentially deadly traffic scenario.
Drivers like Italo Frigoli of Texas report their Teslas ignored flashing lights and lowered crossing gates. In Frigoli’s case, he had to slam the brakes to avoid a collision with an oncoming train. Video reviewed by NBC News confirms his account. When NBC joined him for a follow-up test, the FSD software again failed to detect an approaching train.
NBC spoke with six Tesla drivers who experienced similar issues at rail crossings while using FSD. Four provided video evidence, and the network uncovered at least seven more incidents posted online since mid-2023. Tesla forums also contain widespread written complaints from owners reporting problems at train tracks.
The National Highway Traffic Safety Administration (NHTSA) says it is aware of the issue and is in contact with Tesla, but the company has not responded to media inquiries.
Tesla markets FSD as a driver-assist feature that still requires active human supervision. However, experts warn that the underlying AI operates as a “black box” — even its developers may not fully understand how it makes decisions or why it fails in specific scenarios. One possible explanation is that the system hasn’t been adequately trained on railroad crossings.
Meanwhile, Tesla’s U.S. robotaxi competitor, Waymo, appears to avoid these problems by taking a more cautious approach at crossings and using additional audio sensors to detect trains.
Reported incidents involving FSD include Teslas stopping on tracks, failing to yield to oncoming trains, or even mistakenly turning onto rail lines. In one case in Pennsylvania, a Tesla in FSD mode drove onto train tracks and was hit by a train. No injuries were reported.
Read the full investigation at NBC News.
About Post Author
