Facial recognition tech has outpaced US law – and don't expect the Feds to catch up
Let's be realistic: If the EU can't regulate it well, America definitely won't
Comment If anything could compel the US government to regulate facial-recognition technology, a report sponsored by federal law enforcement urging just that may do the trick.
Produced by a National Academy of Sciences, Engineering, and Medicine (NASEM) committee after spending two years studying the capabilities and implications of the technology at the behest of the FBI and Homeland Security, the report concluded that facial recognition needs to be formally reined in.
According to the panel's write-up, concerns over technical limitations and/or misuse of face recognition systems are well founded and require action from the government to address. As part of its conclusions, the committee recommended a presidential executive order to develop guidelines for appropriate use and fresh legislation "to address equity, privacy and civil liberties concerns" regarding the technology.
"It is crucial that governments make tackling these issues a priority," said Jennifer Mnookin, co-chair of the committee behind the report. "Failing or choosing not to adopt policies and regulations … would effectively cede decision-making and rulemaking on these important questions of great public concern entirely to the private sector and the marketplace."
The report backs up a lot of what's been previously established about facial recognition technology, namely that it has issues with bias across racial and gender demographics – such as wrongly identifying women or people of color – and this has led to some bans on its use.
False negative rates, the report found, were higher by a factor of three for women and non-whites, with the authors citing "algorithms designed in Western countries and trained mostly on White faces" as the culprit.
"Much progress has been made in recent years to characterize, understand, and mitigate phenotypical disparities in the accuracy of FRT [facial-recognition technology] results," the report found, while also noting that "these performance differentials have not been entirely eliminated."
While those facts aren't anything new, it's nice to know Homeland Security and the FBI are finally being told this by a report they had a hand in (it was made independently of the agencies, but they did provide guidance on some of the issues investigated) - but will the findings stick?
There's an increasing number of federal agencies using face recognition systems, and American states are adopting the technology at increasing rates as well, with few legal limits stopping them from implementing it any way they see fit.
It will come as no surprise to US government watchers that there's next to no federal regulation governing the appropriate use of facial recognition technology - a fact that the report takes pains to spell out as a serious problem.
"Facial recognition technology has the potential to impact civil liberties, human rights, and privacy in meaningful ways, because it changes the scale and cost of collecting detailed data about a person's every move," said Edward Felten, another co-chair of the committee behind the report. "The number of uses will continue to expand as the technology becomes more widespread and inexpensive."
Rules don't mean protection
It would be nice to think that US regulation of face recognition would necessarily lead to civil rights protections and limitations on its use, but we need only look to the European Union to see that's not the case.
It's been a month since EU Parliament and Council agreed on language for the bloc's AI Act that carved out exceptions for law enforcement use of "remote biometric identification," (ie, face recognition) that EU parliamentarian Patrick Breyer said opens the door for mass surveillance across the EU.
"It appears the EU intends to compete with China not only technologically but also in terms of high-tech repression," Breyer, the German Pirate Party's sole MEP, said in a statement.
Breyer said the AI Act's face recognition carveouts could allow EU citizens to be tracked and arrested for petty offenses, like instances of the ousting of homeless people in Italy justified under trespassing rules, and the monitoring of political demonstrators at the last G20 summit.
- China – which surveils everyone everywhere – floats facial recognition rules
- Senate bill aims to stop Uncle Sam using facial recognition at airports
- Who in America is standing up to privacy-bothering facial-recognition tech? Maine is right now leading the pack
- Get a $25 gift card if you help the US check whether these facial logins really work
Breyer wasn't the only one hitting out at the facial recognition allowances included in the EU's AI Act - Amnesty International also called language in the Act "a hugely missed opportunity to stop and prevent colossal damage to human rights."
Citing the original language in the proposal (removed in negotiations last month) that included an unconditional ban on live face recognition, Amnesty International AI advocacy advisor Mher Hakobyan said EU regulators have set a "devastating" precedent for other nations to follow.
"While proponents argue that the draft allows only limited use of facial recognition and subject to safeguards, Amnesty's research … demonstrates that no safeguards can prevent the human rights harms that facial recognition inflicts, which is why an outright ban is needed," Hakobyan said.
Returning to the United States, let's be frank: Congress isn't going to be harsher on tech regulations than the European Union, and the corporations behind facial recognition technology have already taken steps to ensure they won't be held back.
Per Open Secrets, the face recognition lobbying game is already in full swing with more than 32 companies having lobbied Congressional representatives for favorable laws in 2021. That lobbying has continued unabated.
Some states, like Massachusetts, have attempted to take action to curb face recognition within their own borders. But, according to 2020 GAO data shows more than half of US states have already purchased their own face recognition systems and allowed federal agencies to have access.
In other words, it's already here, and it's pretty widespread.
As for the White House, whether President Biden will issue an executive order addressing the NASEM report's recommendations is unknown - we asked, but didn't hear back. Then again, even if Biden takes action there's nothing to stop another administration from simply undoing his executive order and pursuing a different agenda.
Not to be pessimistic, but the likelihood the US federal government passes comprehensive regulations on the use of facial recognition seems infinitesimally low, even if the FBI and DHS are being told by their own sponsored report that it needs to happen as soon as possible.