San Francisco is first US city to ban facial recognition

[zenkai]

Source

Legislators in San Francisco have voted to ban the use of facial recognition, the first US city to do so.

The emerging technology will not be allowed to be used by local agencies, such as the city’s transport authority, or law enforcement.

Additionally, any plans to buy any kind of new surveillance technology must now be approved by city administrators.

Opponents of the measure said it will put people’s safety at risk and hinder efforts to fight crime.

Those in favour of the move said the technology as it exists today is unreliable, and represented an unnecessary infringement on people’s privacy and liberty.

In particular, opponents argued the systems are error prone, particularly when dealing with women or people with darker skin.

"With this vote, San Francisco has declared that face surveillance technology is incompatible with a healthy democracy and that residents deserve a voice in decisions about high-tech surveillance," said Matt Cagle from the American Civil Liberties Union in Northern California.

"We applaud the city for listening to the community, and leading the way forward with this crucial legislation. Other cities should take note and set up similar safeguards to protect people's safety and civil rights."

The vote was passed by San Francisco’s supervisors 8-1, with two absentees. The measure is expected to be officially passed into city law after a second vote next week.
Image caption The move angered campaigners who said the tech would help fight crime

"Instead of an outright ban, we believe a moratorium would have been more appropriate," said Joel Engardio, vice-president of Stop Crime SF.

"We agree there are problems with facial recognition ID technology and it should not be used today. But the technology will improve and it could be a useful tool for public safety when used responsibly. We should keep the door open for that possibility."

The new rules will not apply to security measures at San Francisco’s airport or sea port, as they are run by federal, not local, agencies.

Some campaigners unsuccessfully urged for the measures not to apply to local police. While San Francisco’s officers do not currently use facial recognition technology, a number of other police forces across the US do.

Do you believe this is good or bad and why?

[Independent voter]

Well, good for them. I don't like the idea of the state knowing I was walking down 7th Ave on 23rd of May, 2019 at 6:31PM. The state can mind its own business.

[Lanfall]

They just didn't want to have to give citations to all the people shitting in the streets once they recognized them.

[Wewlad]

Sometimes a step back is good.

[Donatello Trumpi]

Do you believe this is good or bad and why?

I am not a fan of face recognition, but knowing San Franshitshow, they probably banned it because 'IT'S RAYCISS' or some shit along the line.

[Jonnusthegreat]

This sounds dumb. Isn't SF like, the tech capital of the world right now?

[Lanfall]

I am not a fan of face recognition, but knowing San Franshitshow, they probably banned it because 'IT'S RAYCISS' or some shit along the line.

Maybe it was making ICE's job to easy?

[Elegiac]

I am not a fan of face recognition, but knowing San Franshitshow, they probably banned it because 'IT'S RAYCISS' or some shit along the line.

That's precisely it, actually; it turns out having a male dominated IT industry tends to produce facial recognition algorithms that have trouble distinguishing between individual women or people of colour for....some reason...

[Donatello Trumpi]

That's precisely it, actually; it turns out having a male dominated IT industry tends to produce facial recognition algorithms that have trouble distinguishing between individual women or people of colour for....some reason...

oh for fucks sake...

[Elegiac]

oh for fucks sake...

Garbage in, garbage out.

New research out of MIT’s Media Lab is underscoring what other experts have reported or at least suspected before: facial recognition technology is subject to biases based on the data sets provided and the conditions in which algorithms are created.

Joy Buolamwini, a researcher at the MIT Media Lab, recently built a dataset of 1,270 faces, using the faces of politicians, selected based on their country’s rankings for gender parity (in other words, having a significant number of women in public office). Buolamwini then tested the accuracy of three facial recognition systems: those made by Microsoft, IBM, and Megvii of China. The results, which were originally reported in The New York Times, showed inaccuracies in gender identification dependent on a person’s skin color.

Gender was misidentified in less than one percent of lighter-skinned males; in up to seven percent of lighter-skinned females; up to 12 percent of darker-skinned males; and up to 35 percent in darker-skinned females.

“Overall, male subjects were more accurately classified than female subjects replicating previous findings (Ngan et al., 2015), and lighter subjects were more accurately classified than darker individuals,” Buolamwini wrote in a paper about her findings, which was co-authored by Timnit Gebru, a Microsoft researcher. “An intersectional breakdown reveals that all classifiers performed worst on darker female subjects.”

[Celista]

ITT: 90% of respondents support the ban, yet still find a reason to shit on SF.

[Independent voter]

ITT: 90% of respondents support the ban, yet still find a reason to shit on SF.

That's a bad choice of words, right there.

[Zython]

Maybe it was making ICE's job to easy?

Why are those bad reasons for banning it?

- - - Updated - - -

oh for fucks sake...

Well, why DOES it have problems with those people?

- - - Updated - - -

They just didn't want to have to give citations to all the people shitting in the streets once they recognized them.

Yes, Mr. Lanfall. People are homeless, and this makes you the victim.

[PC2]

Facial recognition works the best when everybody in the country has to get their faces scanned periodically. This would allow law enforcement to identify criminals who don't have a prior record, or even people who shouldn't be here if they don't have a file. Whether it's a good idea depends on how much we trust the government to only use the information for legitimate purposes.

- - - Updated - - -

That's precisely it, actually; it turns out having a male dominated IT industry tends to produce facial recognition algorithms that have trouble distinguishing between individual women or people of colour for....some reason...

No that can't be true if they are using ML. Programmers and engineers don't have any control over the patterns that are learned. Any bias would be entirely on the *training data*. The data gatherers don't necessarily have to be IT professionals.

- - - Updated - - -

Garbage in, garbage out.

You realize that your argument is discriminatory towards white guys, right? Being less identifiable actually gets those people off the hook. Being more identifiable is not advantageous for the people who violate the law.

[nymphetsss]

Score one for the bad guys

[zenkai]

That's a bad choice of words, right there.

This made me LOL.

[Elegiac]

You realize that your argument is discriminatory towards white guys, right? Being less identifiable actually gets those people off the hook. Being more identifiable is not advantageous for the people who violate the law.

Yeah, being much less likely to register as a false positive is so discriminatory. /s

It also ignores the point; that these systems are subject to bias and are therefore not justifiable due to how intrusive they are.

[zenkai]

I am very conflicted on this. On one hand, I am enthusiastically, 100% for technology and progress. Face recognition is hardly as unreliable as the dudes in the article are trying to make it. Some cheap, old, trash systems might be unreliable, new stuff is much better, yet obviously not perfect - nothing is. Not only that, but machine learning can be used to improve accuracy, these systems will get better and better.

On the other hand, I am and always was a privacy advocate. Having cameras everywhere watching us is the opposite of privacy. It's nightmarish. But... it's hard not to notice that we're always under surveillance, at least those of us that are taking a relaxed, normal approach for accessing the web. One would have to go through a lot of inconvenient steps to hide their presence online.

Overall, I'd reluctantly say this is a good thing. But I am sure that in the end, we'll still get to 100% surveillance coverage in the civilized areas of the world.

I kind of feel the same way.

[Pooti]

Yeah, being much less likely to register as a false positive is so discriminatory. /s

It also ignores the point; that these systems are subject to bias and are therefore not justifiable due to how intrusive they are.

Are they less biased than the current identification measures, such as eye-witness accounts?

[Nfinitii]

Idk if I would ban it, it could be a useful tool to help catch criminals. Obviously b/c of its inaccuracies it should be backed up with other evidence or used as a starting point for investigation.