Executive Summary
This report examines how the transition toward Software-Defined Vehicles (SDVs)
and the growing adoption of autonomous driving and ADAS technologies are shifting
automotive safety requirements from hardware-centered control to software- and
data-driven risk management. As these systems expand in scope and complexity,
software functional safety is being reorganized to address perception limits, human
misuse, algorithmic uncertainty, and system-level interactions.
The study analyzes Tesla’s Autopilot as a representative Level 2 driver-assistance
system, focusing on the 2019 rear-end collision in Florida and the subsequent 2025
verdict by the U.S. District Court for the Southern District of Florida (Barrett v.
Tesla, Inc., Case No. 1:21-cv-21940-BB). The Autopilot system failed to detect a
stationary vehicle and the AEB did not activate, while the driver was momentarily
inattentive, resulting in the death and injury of third-party pedestrians. The jury
assigned 67% negligence to the driver and 33% to Tesla, and imposed punitive
damages due to concerns regarding withheld crash data and misleading marketing.
This case is significant as one of the first to affirm that even SAE Level 2 systems
impose obligations on manufacturers regarding foreseeable misuse, adequate
warnings, and operational transparency.
The report interprets the case findings through international standards including
ISO 26262 (Functional Safety), ISO 21448 (SOTIF), and ISO/PAS 8800 (AI Safety). ISO
26262 focuses on E/E system faults, SOTIF addresses risks that arise from
functional insufficiencies without faults, and ISO/PAS 8800 extends safety
management to AI/ML-based perception and decision models. The analysis also
references relevant global regulations such as UNECE R157 (ALKS) and R171 (DCAS).
Overall, the Autopilot verdict highlights that software safety has become a legal
and regulatory issue linked to product liability, punitive damages, data transparency,
and AI governance. As digital transformation accelerates in safety-critical sectors, it
is essential to establish baseline software safety requirements?such as Korea’s
“Guidelines for Assuring Software Safety” and to advance integrated Safety-Security
governance frameworks that support emerging digital risks.