AI Self-Driving Cars Can Be Stubborn
In this 6‑minute episode, Dr. Lance Elliott explores the concept of "stubbornness" in self‑driving cars, using a tow‑truck flatbed scenario to illustrate how AI may refuse or fail to perform unexpected maneuvers. He explains that current autonomous systems are optimized for routine point‑to‑point travel and lack programming for edge cases like climbing a ramp onto a tow truck, leading to a perceived refusal similar to a mule’s obstinacy. Elliott discusses potential solutions, such as software updates from fleet operators or remote human control, while emphasizing that the AI isn’t sentient—it simply lacks the necessary data and instructions. He concludes that as AI improves, these corner‑case refusals should diminish, but for now, users must understand the limitations of autonomous driving.
AI Self-Driving Cars Might Simply Be Robots That Drive
In this 7‑minute episode, Dr. Lance Elliott explores the concept of a "driving robot" that can sit in a conventional car and operate it using standard controls, as an alternative to embedding autonomous technology directly into the vehicle. He outlines...
AI Self-Driving Cars Dealing With Roadway Zipper Merging
In this episode, Dr. Lance Elliott explores the contentious practice of zipper merging and how it challenges both human drivers and emerging AI-driven autonomous vehicles. He outlines the two main driver philosophies—early merge versus late merge—highlighting the safety and efficiency...
AI Self-Driving Cars That Know When They've Been In A Car Crash
In this episode, Dr. Lance Elliott explores whether AI-driven Level 4 and Level 5 self‑driving cars can recognize that they have been involved in a crash. He explains that unlike human drivers, autonomous systems lack bodily sensations, so crash detection...