What We’re Reading
Tesla launches ‘actually smart summon’ or ‘ass’ (Electrek, September 3, 2024)
Tesla has introduced the long-awaited Actually Smart Summon (ASS) feature, alongside a simpler Dumb Summon option, in its latest software update. Actually Smart Summon enables Tesla owners to call their vehicle from a parking spot to a chosen nearby location via the Tesla App. The release notes humorously describe ASS as a magical tech experience, allowing your car to autonomously navigate to your position or destination.
Dumb Summon offers basic functionality, allowing the car to move forward or backward with app-based controls. To use these features, owners simply access the Summon tab in the Tesla app, gaining live camera feeds to oversee the vehicle’s movements. While employing these features, vigilant supervision is advised, particularly in parking scenarios. Currently, ASS is available for cars equipped with Hardware 4, with future availability for Hardware 3 suggested imminently.
Woman using hands-free driving feature hits car, killing 2 in Pennsylvania, cops say (Miami Herald, September 3, 2024)
A Pennsylvania woman, Dimple Patel, turned herself in on Tuesday this week after hitting and killing two men while driving her Ford Mustang Mach-E in the vehicle’s hands-free driving mode back in March of this year. At around 3 AM on March 3 in Philadelphia, two drivers were pulled over on Interstate 95 when Patel’s yellow Mach-E slammed into a vehicle from behind, killing both men. The Mach-E is reported to have been traveling at around 70 miles per hour with BlueCruise hands-free driving feature enabled, as well as Adaptive Cruise Control when the crash occurred. Ford has commented on the incident, stating, “There is no priority higher than safety at Ford, and we have been collaborating fully with all relevant organizations on the investigations into these incidents.”
The marketing materials for Ford’s BlueCruise state that drivers should continue to keep their eyes on the road when BlueCruise is enabled. Pennsylvania State Police urge drivers to “be aware and ready to regain control if the situation warrants your immediate reaction.”
Researchers probe safety of AI in driverless cars, find vulnerabilities (Tech Xplore, September 2, 2024)
Artificial intelligence (AI) is a critical component in self-driving vehicles, facilitating decision-making, sensing, and predictive modeling. However, its vulnerability to attacks is a concern being explored by researchers at the University at Buffalo. Their studies indicate that malicious actors could potentially manipulate these AI systems, causing them to fail. For instance, 3D-printed objects might camouflage a vehicle from AI radar systems, rendering it undetectable.
Although current autonomous vehicles remain safe, this research could impact the automotive and tech industries, and influence regulatory policies. Tests conducted at UB have shown that the most commonly used radar systems for object detection, though reliable, can be compromised using strategically placed “tile masks.” These discoveries underline potential attack motives such as insurance fraud or competition, and highlight the gap in security measures, which researchers aim to address by developing robust defenses.