Santa Clara, California — With artificial intelligence (AI) becoming more prevalent in the automotive industry, the need for diversity and inclusion it becoming more apparent.
Neda Cvijetic, the senior manager of autonomous vehicles and computer vision at Nvidia, says that the industry has been using artificial intelligence to improve systems but because of biases the vehicles may not be as safe for women and children as they are for men.
“What ended up happening is that these safety settings, crash settings, airbag settings, were systematically suboptimally set for these other types of occupants because of the bias that was there in the design of the problem,” said Cvijetic.
Santa Clara, California — With artificial intelligence (AI) becoming more prevalent in the automotive industry, the need for diversity and inclusion it becoming more apparent.
Neda Cvijetic, the senior manager of autonomous vehicles and computer vision at Nvidia, says that the industry has been using artificial intelligence to improve systems but because of biases the vehicles may not be as safe for women and children as they are for men.
“What ended up happening is that these safety settings, crash settings, airbag settings, were systematically suboptimally set for these other types of occupants because of the bias that was there in the design of the problem,” said Cvijetic.
Aside from crash test dummies, there are many other ways a vehicle can be programmed to have biases. One of the major problems that autonomous vehicles are going to face is recognizing different skin tones while on the road. According to a 2019 Georgia Institute of Technology Study, automated crash-avoidance systems were less likely to be able to sense a pedestrian with darker skin tones.
“It’s very critical for these autonomous vehicles to be able to recognize humans both inside of the car and outside the car,” said Cvijetic. “There needs to be data for the car to be able to recognize pedestrians, all kinds of people across different genders, racial backgrounds, age groups. … We cannot bias a system towards detecting one type of pedestrian but not another type of pedestrian.”
Ways to combat biases in artificial intelligence systems is to make sure the team developing the system is diverse, this means having all genders, skin tones and different age groups involved in the process because they will all have different needs and input.
Indu Vijayan, director of product management at lidar-maker AEye Inc., says that diversity is key to creating safe AI systems for everyone.
“Different people have different ideas,” she said. “Bringing them together will give you a more holistic approach, thereby reducing any kind of biases in the system, which inadvertently will enable the system to be safer and to take the right action when things happen.”
