In early September, Portland, Oregon city council met virtually to consider sweeping laws banning the use of facial recognition technology. The bills would not only prevent the police from using it to expose demonstrators and people who were being held in surveillance images. You would also prevent corporations and a host of other organizations from using the software to identify an unknown person.
During the public comment time, a local, Christopher Howell, said he had concerns about a blanket ban. He gave a surprising reason.
"I'm involved in the development of facial recognition, which will actually be used by Portland police officers because they don't identify themselves to the public," Howell said. During the summer when the city was seized by anti-police violence demonstrations, the heads of the department told uniformed officers that they could tape their names. Mr. Howell wanted to know: Would his use of facial recognition technology become illegal?
Portland Mayor Ted Wheeler told Mr. Howell his project was "a little scary," but a city attorney clarified the bills would not apply to individuals. The council then passed the law unanimously.
Mr. Howell was offended by Mr. Wheeler's characterization of his project, but relieved that he could keep working on it. "There's a lot of excessive violence here in Portland," he said in a telephone interview. "Knowing who the officers are seems like a baseline."
Mr Howell, 42, is a lifelong protester and self-taught. In graduate school, he began working with neural network technology, an artificial intelligence that learns to make decisions from data it provides, such as: B. Pictures. He said police gassed him tearfully during a lunchtime protest in June and began researching how to build a facial recognition product that could undo officials' attempts to protect their identities.
"It was a kind of 'shower thought' for me and just a kind of intersection between what I know to do and what my current interests are," he said. “Accountability is important. We need to know who is doing what so that we can deal with it. "
Mr. Howell is not alone in his pursuit. Law enforcement agencies have used facial recognition to identify criminals using photos from government databases or from the public internet through a company called Clearview AI. But now activists around the world are turning the process around and developing tools to expose law enforcement in cases of wrongdoing.
"It doesn't surprise me in the slightest," said Clare Garvie, an attorney at Georgetown University's Center on Privacy and Technology. "I think some people will say, 'All is fair in love and war,' but it shows the risk of developing this technology without thinking about its use in the hands of all possible actors."
The authorities mentioned so far were not pleased. The New York Times reported in July 2019 that Colin Cheung, a protester in Hong Kong, had developed a tool to identify police officers from online photos. After posting a video about the project on Facebook, he was arrested. Mr. Cheung eventually gave up the job.
This month, artist Paolo Cirio posted photos of 4,000 French police officers' faces online for an exhibition called Capture, which he described as the first step in developing a face recognition app. He collected the faces from 1,000 photos he had collected from the Internet and photographers who took part in protests in France. Cirio, 41, took down the photos after the French interior minister threatened legal action but said he hoped to republish them.
"It's about everyone's privacy," said Cirio, who believes facial recognition should be banned. "It is childish to try and stop me as an artist who is trying to address the problem instead of addressing the problem itself."
Many police officers around the world cover their faces in whole or in part, as recorded in recent videos of police violence in Belarus. Last month, Andrew Maximov, a technologist from the country who now lives in Los Angeles, uploaded a video to YouTube showing how facial recognition technology can be used to digitally remove the masks.
In the simulated footage, the software compares masked officers with full images of officers taken from social media channels. The two images are then merged so that the officers are shown in uniform and their faces are displayed. It is unclear whether the matches are correct. The video, previously reported by a news site about Russia called Meduza, has been viewed more than a million times.
"For a while everyone was aware that the big boys could use this to identify and suppress the little boys, but we are now approaching the technological threshold where the little boys can do it to the big boys," he said Maximov, 30, said. "It's not just the loss of anonymity. It's the risk of shame."
These activists say it has become relatively easy to develop facial recognition tools thanks to the standard image recognition software available in recent years. In Portland, Mr. Howell used the TensorFlow platform provided by Google, which enables people to create machine learning models.
"The technical process – I'm not inventing anything new," he said. "The big problem here is getting high quality images."
Mr Howell collected thousands of pictures of Portland police officers from news articles and social media after finding their names on the city's websites. He also requested a list of police officers with their names and personnel numbers in public records, but it was denied.
Facebook was a particularly helpful source for pictures. "Here they are all having a barbecue or whatever, sometimes in uniform," said Mr. Howell. "There are few people I can reasonably do as an individual."
Mr Howell said his tool was still in the works and could only detect about 20 percent of the Portland police. He didn't make it public, but he said it already helped a friend confirm an official's identity. He declined to give further details.
Derek Carmon, an information officer for the Portland Police Bureau, said that "name tags have been changed to personnel numbers during protests to prevent officers from being doxxing," but that officers are required to wear name tags for "non-protest" duties. " People could file complaints using an officer’s ID number who declined to comment on Mr. Howell’s software.
Older attempts to identify police officers have relied on crowdsourcing. The news service ProPublica is asking readers to identify officers in a series of videos about police violence. In 2016, an anti-surveillance group in Chicago, Lucy Parsons Lab, launched OpenOversight, a "publicly searchable database of law enforcement officers." It urges people to upload photos of uniformed officers and match them with the officers' names or ID numbers.
“We have been careful about what information we request. We don't want to encourage people to follow officers with their children into playgrounds, ”said Jennifer Helsby, lead developer of OpenOversight. "It has resulted in officers being identified."
For example, the database helped journalists with the Invisible Institute, a local news organization, identify Chicago officials who beat protesters with batons this summer, according to the institute's director of public strategy, Maira Khwaja.
Photos of more than 1,000 officials have been uploaded to the website, Ms Helsby said. Versions of the open source database have been launched in other cities, including Portland. This version is called Cops.Photo and is one of the places Mr. Howell obtained identified photos from police officers.
Mr Howell originally wanted his work to be public, but is now concerned that giving his tool to others would be illegal under the city's new facial recognition laws, he said.
"I got legal counsel and will be catching up more," Howell said. He described it as "unwise" to publish an illegal facial recognition app because the police "will not appreciate it from the start".
"I would be naive if I didn't worry about that," he added. "But I think it's worth it."