Individuals with autism often struggle with identifying non-verbal cues that neurotypical people recognize from an early age. With the help of facial expression recognition (FER), it is possible to develop an AI program that associates simple cues such as emojis with people's emotions. This program aids individuals with autism in learning and identifying non-verbal cues to better interpret the world around them.
It is known that FER more often misinterprets the emotions of people of colour (POC) in comparison to Caucasians. To optimize accessibility and inclusivity, it is important to minimize any bias in the data by developing more data associated with POC.
The overall goal is to minimize any bias within FER technology in order to provide AI technology that can aid individuals with autism identify the emotions of other individuals, regardless of their race with maximum accuracy.