AUDIE CORNISH, HOST:
Now we’re going to look more broadly at what’s been revealed today about ICE turning to DMV offices for help with facial recognition – that is, using driver’s license photographs and algorithms to identify people suspected of being in the country illegally.
Now, this collaboration was unearthed by a team at Georgetown University, and here to brief us is NPR’s Aarti Shahani. Hey there, Aarti.
AARTI SHAHANI, BYLINE: Hi.
CORNISH: I understand that in the past, ICE has gone to DMV offices and just asked for records on immigrants. We just heard about the case in Vermont that alleges that much. What exactly is new here?
SHAHANI: So what was uncovered is that ICE agents, while looking for undocumented people, ended up having extraordinary access to the state records of American citizens. Lawyers at Georgetown’s Center on Privacy and Technology have been submitting Freedom of Information lawsuits to DMVs around the country, trying to learn what they can about how each state does or doesn’t collaborate with ICE.
Three states – Utah, Washington and Vermont – handed over documents showing that ICE was not just reaching out to them with targeted searches. ICE agents were not just saying, hey, here’s a specific person we want, their full name and date of birth; can you share what you’ve got on them?
Instead, ICE was saying, hey, we’ve got a high-resolution picture of somebody who entered the U.S. on a visa. We believe this person overstayed. Can you take this picture, run it through your database which includes many, if not mostly, U.S. citizens and give us the faces that match this one? Basically, help us pair of visa photos to license photos. These requests happen from 2015 to 2017.
CORNISH: And I understand according to the Georgetown findings, these three states did do it – right? – hand over facial recognition matches to ICE.
SHAHANI: Well, according to the FOIA documents, Utah and Vermont did, and with Washington, it’s unclear. The agency told The Washington Post, which had first reported this, that they just respond to court orders. As you’d mentioned, NPR also reached out to ICE. And I’d add that ICE – the ICE spokesperson said that what they’re doing is consistent with what other law enforcement agencies do.
Now, it is important to point out facial recognition has done plenty of good in this world. It’s helped find missing children and reunite with them with their families. But in this instance, activists have raised concerns. They say there is a bait-and-switch going on. Not every state lets undocumented immigrants get a license. These three states are among those that do. They’re signaling to undocumented people it is safe to come here and apply for your driver’s license. But then the DMVs are turning around and handing files over to deportation officers.
CORNISH: What about the reliability of the technology itself? Is facial recognition far enough along that we can be counting on it in this way?
SHAHANI: Well, last year, the MIT Media Lab did a study. It found that leading software was accurate 99% of the time when it came to identifying the gender of white males. So failure in only 1 out of a hundred. But with darker-skinned women, it failed to identify them as women 1 out of 3 times. So that is a huge disparity. Alvaro Bedoya, one of the Georgetown lawyers – he called the ICE-DMV tag-teaming a dragnet. And he says citizens of color are particularly vulnerable here. This is him.
ALVARO BEDOYA: The question that people need to ask themselves in these states is not, am I my undocumented, but rather, does a flawed face recognition algorithm think that I look like someone who is undocumented?
SHAHANI: And he is not alone in his concerns. San Francisco recently banned the use of facial recognition by police and city agencies. The company that’s the largest maker of police body cameras says it’s not going to sell facial recognition tech for now because it’s just not reliable enough. Experts from some of the biggest tech companies like Google and Microsoft have petitioned Amazon to stop selling facial recognition tech for that same reason, though Amazon is still selling it, including to government agencies. So we’re seeing very powerful entities at odds.
CORNISH: That’s NPR’s Aarti Shahani. Aarti, thank you.
SHAHANI: Thank you.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.