March 19 (Reuters) – The U.S. auto safety regulator has deepened a probe into Tesla’s driver-assistance system to evaluate its effectiveness ‌in identifying degraded road conditions after nine crashes linked to ‌the technology including one that was fatal.

The probe covers roughly 3.2 million vehicles, across ​multiple models, accounting for nearly all vehicles the company has sold in the United States.

The move marks a significant escalation that could lead to a recall or other enforcement action if regulators find a safety defect.

Tesla’s ‌ambitions to deploy fully ⁠autonomous vehicles and robotaxis hinge on the reliability and safety of its Full Self-Driving (FSD) system, which remains under ⁠regulatory scrutiny.

The probe focuses on Tesla’s “degradation detection” system, which is designed to recognize when visibility is impaired and warn drivers to take control.

Available data raised ​concerns ​that Tesla’s degradation detection system, both ​in its original form and ‌after updates, failed to detect reduced visibility or adequately warn drivers in situations such as glare or other obstructions, NHTSA said.

Tesla did not immediately respond to a Reuters request for comment.

The regulator said it had identified nine crashes linked to the issue, two of which involved ‌injuries.

Tesla’s post-incident analysis indicated that a software ​update to the degradation detection system, if ​installed at the time, may ​have affected three of the nine incidents, the ‌regulator said.

In crashes reviewed by regulators, ​the system did ​not recognize conditions that impaired camera visibility or issue alerts until just before impact.

Regulators also identified additional crashes in similar environments where ​the system failed ‌to detect lower visibility or did not provide drivers with ​sufficient time to respond.

(Reporting by Akash Sriram in Bengaluru; Editing ​by Shinjini Ganguli and Anil D’Silva)