The Computing Technology Association (ACM) – the largest and oldest international organization in the computer field – has asked lawmakers to immediately stop using face recognition technology by government agencies and private companies.
The reason is the documented ethnic, racial and gender bias that these systems exhibit. In a letter published by the US Technology Policy Committee (USTPC), a group of authors acknowledges that this technology will become better in the future, but so far it is not mature enough and threatens human rights, VentureBeat reports.
“The consequences of such bias,” the statement says, “often go beyond inconvenience and lead to serious damage, in particular, to life, property status and fundamental rights of individuals from certain demographic groups, including the most vulnerable segments of the population.”
In the past, less numerous and influential organizations such as the American Civil Liberties Union or the Algorithmic Justice League made such statements, but for the first time an association uniting nearly 100,000 computer technology professionals came out in support of the ban on face recognition technology, HighTech + reports.
One of the proofs of the technology’s malfunctioning is the statement by James Craig, the head of Detroit police (USA), that the program that recognized faces was mistaken in 96%. “If we relied solely on the program, we would not have disclosed from 95 to 97% of cases,” he said.
In addition, the authors listed the principles of ethical face recognition in a letter. These include:
1. The face recognition system should undergo an independent audit and a thorough review by the government.
2. People should be made aware that such technology is being used.
3. Organizations using this technology should be held accountable for harm caused by it.
It should be emphasized that the appeal does not call for a permanent ban on face recognition, but it requires a temporary moratorium, while the standards of accuracy, as well as laws and regulations, mature.