Unveiling Widespread Exclusion in AI Facial Recognition Systems: A Perspective from the Disfigurement Charity
In the digital age, facial recognition technology has become a common feature in countless daily activities, from unlocking phones to navigating social media. However, for the 72 million individuals worldwide living with a facial difference, these tools often lead to widespread inconvenience and frustration. A recent campaign, "Visibly Invisible," launched by Face Equality International and its global alliance of non-profits, aims to bring attention to this issue.
The "Visibly Invisible" campaign is supported by survey research involving individuals with facial differences caused by scarring, burns, reconstructive surgery, craniofacial conditions, and various other congenital malformations. Although the sample size is relatively small at 60 people, the insights gained offer a glimpse into this often overlooked topic.
Survey respondents reported numerous challenges when interacting with facial recognition technology. For instance, 32% of users experienced issues uploading photos to social media sites, while only 21% of those with facial differences felt banking apps worked for them every time. Many also reported difficulties using electronic passport gates and unlocking their phones with facial recognition.
Phyllida Swift, CEO of Face Equality International, commented on the research, saying, "Facial/visible difference is consistently an afterthought in the policies that should be designed to protect communities from overt marginalization." Actors living with facial differences, like Adam Pearson, have also shared their experiences, highlighting the discrepancy between a fully recognized face in movie screens and billboards and the struggles with facial recognition technology.
To tackle these challenges, Face Equality International calls for more research into AI-powered facial recognition tools. This research should aim to understand failure points and other technological glitches, as well as how the tools behave with individuals without facial differences for comparison. Additionally, improvements in ethical considerations and bias reduction are essential to ensure fairness in facial recognition systems.
Another critical hurdle is the legislative framework, as facial differences have traditionally been overlooked in both diversity and disability inclusion initiatives. By improving these tools and addressing legal protections, a more inclusive and equitable future for individuals with facial differences can be realized. While resolving these challenges may not occur overnight, the "Visibly Invisible" campaign serves as a vital step towards highlighting these important issues.
The campaign, "Visibly Invisible," has gained support on platforms like Facebook and Instagram, mobilizing advocates worldwide. This digital mobilization also extends to TikTok, where users with facial differences share their experiences and promote the campaign's message.
However, the lack of diversity in AI development teams can contribute to biases in facial recognition technology, leading to further marginalization of individuals with facial differences. For instance, a person with neurofibromatosis might face challenges with facial recognition systems that struggle to process unusual facial features.
Moreover, some individuals with facial differences may not qualify for creditlines due to inconsistencies in facial recognition data, exacerbating financial marginalization. Addressing this issue requires collaborative efforts between tech companies, policymakers, and non-profits like Face Equality International.
The unique identifier 'f60ebcd5d5b002c95c435bc173828cb1' assigned to users by Face Equality International is designed to circumvent facial recognition challenges, providing a workaround for individuals with facial differences. This innovative approach highlights the organization's commitment to promoting face equality and overcoming barriers in the digital age.
