Why are driverless cars still so bad at driving? | Jennifer Dukarski | TEDxDetroit

Autonomous Vehicle Technology & Inclusivity - The core issue is the need for inclusive design in autonomous vehicle (AV) technology to ensure equitable outcomes for all individuals, including those with disabilities, the elderly, and underserved communities [1] - Current AV systems, exemplified by Tesla's full self-drive, rely on data collected from a demographic that is primarily high-income, male, and white, which introduces bias [1] - Data from Lending Tree indicates that Tesla drivers are the second most likely to have vehicle incidents and the most likely to have vehicle crashes per thousand vehicles [1] - A 2019 Georgia Tech study found a nearly 10% difference in the ability of vision systems to detect dark-skinned people versus light-skinned people [1] - A 2023 King's College study showed improvement with a 7% difference but also revealed that these systems are 20% more likely to detect adults compared to children [1] Inclusion by Design Principles - Inclusion by Design necessitates identifying all people, regardless of gender, age, or race, to ensure AV systems function effectively for everyone [1] - It is crucial to consider the broader community impact, even if the vision system does not directly perceive it, to avoid unintended consequences such as noise pollution from autonomous vehicles [1] - Location matters, and AV systems must be designed to account for diverse environments, whether suburban or urban, and the interactions within those communities [1] Recommendations - Inclusion by Design principles should be integrated into engineering education, considered by standards bodies, and addressed by regulators to advance inclusive mobility [1]