Daniel Castro, vice president of the Information Technology and Innovation Foundation, a Washington think tank funded by the technology industry, says the rules on facial recognition would align with the Biden campaign’s interest in privacy and racial justice. “There was already an opening for possible action, and this could go further with the Biden administration,” he says. “This could be one of the items they are putting on the agenda.”
Biden’s interest in racial justice is in part driven by the protest response to the police murder of George Floyd in Minneapolis earlier this year. The incident prompted IBM to announce in June that it will no longer offer facial recognition. Shortly afterward, Microsoft said it would cease sales to law enforcement until federal legislation was in place. Amazon also stopped law enforcement sales, but only for a year.
Chris Padilla, IBM’s vice president of government and regulatory affairs, said the company was already concerned about inaccuracies in facial recognition, but the Floyd protests had made the company aware of the potential consequences of failure. “We’re not sure these things can work properly enough in a law enforcement context,” he says.
IBM has not commented on a federal ban on law enforcement use, but is pushing for a national discussion on regulatory options. Padilla also says the government should ban exporting the technology to authoritarian countries like China.
“There was already an opening for possible action, and this could go further with the Biden administration.”
Daniel Castro, Vice President of the Foundation for Information Technology and Innovation
Amazon referred WIRED to a 2019 blog post that suggested safeguards for law enforcement use of facial recognition, but also stated that “new technologies should not be banned or condemned for their potential abuse.” A Microsoft spokesman said government rules on whether and how facial recognition should be used “should be based on human rights protections such as privacy, freedom of expression and association.”
Tech companies and some lawmakers broadly agree that facial recognition rules are needed, but there won’t be an easy agreement on how accurate the limits should be. A Washington state law passed in March, backed by Microsoft and introduced by a senator who works for the company, illustrates some of the divisions.
Washington law requires government agencies to disclose information about the use of facial recognition technology and the accuracy of that technology in relation to different populations. It also requires “meaningful human review” when the technology is being used to make critical decisions, and it prohibits law enforcement from using facial algorithms for live video feeds except in an emergency.
Microsoft called the bill an “important model,” but it is more permissive than the government bans on the use of facial recognition in more than a dozen cities, including Boston and San Francisco. Portland passed a law that also bans its use by private companies, against Amazon’s opposition.
The secret of machine learning? Human teachers
Jennifer Lee, project manager for technology and freedom at ACLU in Washington, hopes state law does not become a national model. “We need a strict government ban on surveillance and to make sure agencies and businesses can’t use facial recognition to profile people,” she says. “The approved bill does not contain these measures.” The ACLU is working on a new privacy law for Washington and is hoping to include a requirement that businesses will require consumers to opt out of using the facial recognition technology.
Some proposals published in Congress were stricter than Washington law. In 2019, Senators Roy Blunt, Missouri, and Brian Schatz, D-Hawaii introduced a bill requiring companies to obtain consent before they can collect facial recognition data. Files show that Amazon, IBM and Microsoft have all lobbied lawmakers on the bill. In June, a group of Democratic senators and officials tabled a bill providing a moratorium on the use of biometric technologies, including facial recognition, by federal agencies.
In the absence of federal regulation, face recognition is more accessible and widely used. This, and the breadth of industries advocating this issue in DC, suggest that it will be complicated to work out ground rules. “Face recognition is increasingly the most effective and seamless way to identify people in many different types of applications,” said ITIF’s Castro.