Tesla deploys neural network-powered Full Self-Driving update as NHTSA investigates unexpected lane changes and China prepares conditional approval for domestic rollout.
Tesla's radical rewrite of autonomous driving software sparks both technical praise and safety concerns across three continents.
Neural Network Leap Challenges Industry NormsTesla began deploying its Full Self-Driving v12.3 software update on 22 May 2024, replacing 300,000 lines of code with a single neural network trained on 10 million video clips. Early data from LexisNexis Risk Solutions showed 19% fewer collision warnings in urban areas during the first 72 hours of limited North American rollout.
Safety Regulators Launch ProbeThe National Highway Traffic Safety Administration (NHTSA) opened a formal investigation on 28 May 2024 after receiving 112 reports of sudden lane departures across 15 states. Agency documents reviewed by Reuters show particular concern about the system's decision-making transparency.
China's Conditional Approval PathwayTesla submitted required data-security documentation to Chinese regulators on 29 May 2024, seeking approval for domestic FSD deployment. Industry analysts note this comes as Beijing accelerates smart city infrastructure development, with 5G-connected traffic systems now operational in 20 major cities.
Technical Arms Race IntensifiesCEO Elon Musk revealed during Tesla's 24 May shareholder meeting that the company allocated $2 billion for AI training compute in Q2 2024. This investment enabled processing four times more driving scenarios than previous versions, according to internal engineering documents.
Historical Precedents and Regulatory ChallengesThe current NHTSA investigation follows 46 formal probes into Tesla's Autopilot since 2016, including a 2021 examination of phantom braking incidents. Traditional automakers like Ford and GM typically complete 12-18 month validation cycles for advanced driver aids, while Tesla's over-the-air updates sometimes ship within weeks of development.
Automotive safety experts recall how Mobileye's EyeQ4 chip faced similar scrutiny in 2018 when neural network-based pedestrian detection showed inconsistent performance across lighting conditions. The Society of Automotive Engineers notes in its 2024 autonomy report that no Level 3+ system has yet achieved universal regulatory approval without manufacturer liability assumptions.
https://redrobot.online/2025/05/teslas-ai-driven-fsd-v12-3-rollout-faces-regulatory-scrutiny-amid-technical-breakthroughs/
No comments:
Post a Comment