By Isaac Singer '21
A vote to ban the sale of Amazon’s facial recognition software will be conducted on May 22nd, after the Securities and Exchange Commission (SEC) rejected Amazon’s pleas to dissolve the motion on April 3rd. The technology, which is currently being sold to governmental agencies, is said to inaccurately identify select individuals, specifically women and people of color. Amazon shareholders argue that the sale of the technology is contributing to human rights violations and feel that it should not be sold until the technology improves.The inaccurate identification of Black people and Women is no speculation. When the American Civil Liberties Union (ACLU) conducted tests with the recognition technology on members of Congress, 28 individuals were matched to individuals in a database of criminals, with a disproportionate percentage of the 28 being black. Additionally, in January, MIT researchers published a study which revealed that Amazon’s facial recognition technology mistook women for men 19 percent of the time and mistook black women for men 31 percent of the time. In response, Amazon stated that only due to misuse could the technology be harmful, and this is not a legitimate reason to ban the technology. The technology is based on deep learning AI, which improves as it collects more data. The company says the technology has, in reality, been used to fight human trafficking and reunite missing children with parents. When used properly, Amazon argues, the technology has incredible educational and security applications; Amazon executives strongly urge shareholders to vote down the ban.While Amazon clearly has a financial incentive to promote the sale of the technology, they hold a stance which, I believe, is socially responsible. Amazon proposes that their technology should not be self regulated, but rather regulated by the government. They argue that legislation should be enacted which protects civil rights while simultaneously allowing for continued development of the technology. Allowing for companies themselves to decide how they can and should use such technologies is extremely dangerous. Facial Recognition technology is not going away anytime soon. Its ability to identify criminals in crowds, enhanced (and expedited) airport security, and even to find missing people is far too powerful to throw away because of the complex data security and civil rights issues that come along with it. Banning widespread use of such important technology, I believe, is a mistake as it defeats the purpose of the nature of the technology. Machine learning AI improves as it acquires more data so keeping the technology locked away retracts from the value it could bring to society. Regardless of how the vote turns out, the shareholders hold little actual power and it is likely that the sale of facial recognition technology will continue. Sources:https://www.theinquirer.net/inquirer/news/3074213/amazon-shareholders-to-vote-on-if-creepy-ai-face-recognition-should-be-bannedhttps://www.engadget.com/2019/04/15/amazon-shareholder-vote-facial-recognition/https://www.consumeraffairs.com/news/amazon-shareholders-to-vote-to-ban-sales-of-facial-recognition-technology-041519.htmlhttps://www.nytimes.com/2019/04/03/technology/amazon-facial-recognition-technology.htmhttps://venturebeat.com/2019/01/24/amazon-rekognition-bias