Despite efforts to address racism, it remains a persistent problem in our society. Unfortunately, even AI technology has been found to perpetuate discriminatory practices due to the biases of its creators. For example, ChatGPT has displayed evidence of racial bias in its responses to questions. Now, it appears that self-driving car technology may also be susceptible to these same issues.
A recent study conducted by UK and Chinese researchers has revealed that self-driving car technology's pedestrian detection software may not function correctly for individuals with non-white skin tones. Additionally, their findings suggest that self-driving cars may have difficulty detecting children.
One of the self-driving vehicle fears is that it might not work exactly as the theory presents itself. And it looks like this line of thought is true. With the ever-increasing prices of self-driving vehicles, it's a lot of wonder why this technology would have any problems detecting humans at all.
The new study demonstrates how these autonomous cars might find it hard to detect when children or people of color are in traffic. The researchers based their study on race, age, and gender.
To make their findings, they each took turns researching three categories. The first category revealed that pedestrians with darker skin tones may be more difficult to detect, while those with fair skin tones have a better chance of being detected. Additionally, older individuals are more likely to be detected than children. However, the gender category did not provide sufficient results to be taken into consideration.
Jie Zhang, a computer scientist at King's College London and a research team member, discussed the challenges people of color have faced in the past. He also lamented how they now have a new problem of getting hit in traffic due to self-driving car technology.
"Before, minority individuals may have been denied vital services. Now they might face severe injury," Zhang said.
Researchers Comment on the New Study
According to the study, detection systems were 19.67% more likely to detect adults than children. In the same way, those who have fair skin tones stand more than a 7% chance of getting detected than black skin tones. The study further states that it is very important to address the issue of racial bias in pedestrian detection systems before putting them out in self-driving cars. This will be extremely important to reduce any cases of accidents in their vehicles.
"Overall, this study sheds light on the fairness issues faced by existing pedestrian detectors, emphasizing the importance of addressing bias related to age and skin tone. The insights gained can pave the way for more fair and unbiased autonomous driving systems in the future," the study said.
Furthermore, the researchers behind this study are calling on lawmakers to create laws that will regulate the issue of AI and the technologies in self-driving cars. According to them, this prevents loss of lives from pedestrian detection systems not recognizing children and people with dark skin tones.
"It is essential for policymakers to enact laws and regulations that safeguard the rights of all individuals and address these concerns appropriately," the study said.