U.S. TSA expands controversial facial recognition program | Hacker News Reader

U.S. TSA expands controversial facial recognition program

cbsnews.com

77 pointswalterbell26 comments

I don't really get the point of this. I had much longer hair and a full beard when I got my driver's license. I look vastly different now. Every now and again, a TSA agent will casually comment that I don't look anything like my ID photo. They still let me through. Since I was in the active duty military for a long time and able to renew my license from out of state without ever having to go the DMV, before getting my current license, I was as old as 38 using a license with a picture of me taken when I was 23. They still let me through every time.

What is the point of a matching process when failure to match doesn't result in you not being let through?

Like already said Dubai is doing this; additionally Australia has you take a photo at home and load it into an app and then additionally upload your passport to get a visa. My general sense of Australia as a foreigner is that they are very comfortable with watching their citizens and foreigners

That's a shame as some of my favorite products and services such as sublime text and campaign monitor are from Australia.

The stuff that does break out to international scenes is very high quality but politically then somehow have sold themselves to Chinese and Indian businesses and they don't seem to mind being under surveillance.

Wonder if the fact that Australia was a prison island has anything to do with that.

I’ve lived in Sydney most of my life. I suspect that it genuinely is partly because Australia, particularly NSW, did start off as a prison colony, and as a consequence our police force and legislature has always tended towards the conservative and oppressive. NSW is known is Aus as almost always having the most draconian rules for a given thing, compared to other states. NSW is the most populous state and has an outsized influence on federal politics.

Given that we’ve had a conservative (it was the “Liberal” party, but they’re basically like the US republicans) government for most of the last 25 years, it’s not surprising that these factors have come together to make Australia much more of a surveillance state than many of our neighbours.

Interestingly in Dubai International Airport, they have these smart gates that recognize you and re-issue your re-entry eVisa using facial recognition. You walk up, they scan your face and show your name.

No passport gets scanned.

Major US airports have this too for boarding international flights.

For what it’s worth I’ve had no trouble opting out. A polite but firm request to the gate agent did the trick:

“I prefer not to have my face scanned. Will you scan my boarding pass?”

I have also opted out of body scans for TSA (before I pony’d up for Pre-Check) and out of their facial recognition check at Boston Logan.

This is a convenience that will turn into an absolute nightmare for average citizens in a few years time. Imagine saying something the state doesn't like and suddenly half the doors in the country won't open for you.

The last time I re-entered the US, they just had us walk through an area really quickly. I didn't even have to show my passport or any ID. It was kind of nice but also unsettling that they have such a level of confidence in their facial detection.

It’s mostly theater. Someone determined enough can always cross illegally.

The many-eyed monstroisity that sees everyone and is seen by no one keeps growing under the fig leaf of "safety" and "security", fed by people's fear and anxiety.

I don’t think this purely a fear campaign.

Billions are spent on border protection every year for a very broad range of reasons - even before the past 20 years of ‘terrorism bad’. Narcotics, dangerous and malicious persons, people smuggling, customs and international cooperation are legitimate concerns that nations genuinely care are about.

How much money do you think a government contracting company charged for this system?

Considering how crappy and plain the UI is, they must have made a boatload of money for not a very high bar.

The government wants us taking more road trips.

Sadly, no one will blink an eye. I hate to be pessimistic on this, but your average person does not care about their privacy or even most of their rights. I think your typical person will throw just about all of that away for the appearance of safety and security.

If not for job security, surely this is a reason to make software and IT a national trade union.

You can see this at US airports already. Facial recognition is (sometimes? on international flights?) used, but you can opt out. If you do opt out, nothing bad happens. It's not even more inconvenient, the gate agent checks your picture ID instead of the facial recognition box. Still, almost nobody does.

no you are completely misreading it.. the average person is overwhelmed and distracted; uniformed service people and their close allies have no problem with this technology, in fact it fits closely to their existing daily routine and worldview.

Ordinary people are being required to use the same rigor and verification that uniform services use. What is the problem with that? similar fork in the analysis.. from the average person point of view it is invasive and "I didnt agree to this" but that is not what is important IMO; what happens is that the uniform services are prone to deep corruption because the rewards for that, and the ego of that, are just too great. IMO both "I can do this obviously illegal thing since no one can stop me" like some low-level thug mind, and "I make the rules and the rules say I get these privelages" much more common.

This is a disaster for civil society unfolding now. A uniform services takeover of previously civilian and uniform mix. Who runs the ID system? updates the software ? checks for outstanding warrents? etc

The TSA already has your travel data from the airline PNR database [1], which stores your travel data for five years, after which it is moved to cold storage for another ten years. PNR is a contributing system as to how the TSA can provide identity proofing at the checkpoint with only an ID now, versus previously needing your boarding pass [2].

These are governance issues, not technology issues. TSA using facial recognition is no different than an agent comparing you to your ID (except that the automated system increases throughput), depending on what the exception process looks like.

[1] https://www.cbp.gov/travel/clearing-cbp/passenger-name-recor...

[2] https://www.dhs.gov/publication/dhstsapia-018-tsa-secure-fli...

No, a human being comparing your face to a picture is entirely unlike facial recognition. The cop matching your face will not remember you for an arbitrarily long time (or indefinitely), cannot transmit the memory of your face to other cops across space and time, does not catalogue what you are wearing, the delta between your actual face and the image, and all other manner of minutiae.

Cameras and automated systems are likewise fallible in entirely separate dimensions (personal bias, circumstantial factors): the camera does not get tired, but we place a certain unwarranted trust in black-box systems that fail in unexpected ways.

They are absolutely governance issues, but _augmented_ in scale by technological reach.

Don't worry, there will be face scanners for drivers soon, joining the license plate scannners that already exist.

If it’s a diff tool for TSA agents, then sure.

It’s the retention and inability to prove illegal retention that becomes major problems.

Funnily enough your negative comments here regarding this are probably funneled to another program used to rank people by their level of seditiousness

Other data points: - Look-ups of Ted Kaczynski’s manifesto, per his death

- Running his manifesto through OpenAi and Bard to get tl;dr; summary of his points

- Look-ups of Ted Kaczynski, because I can’t spell his name

- This comment

> Privacy advocates worry about the lack of regulations around facial recognition and its tendency to be less accurate with people of color.

What is the cutoff for being less accurate? What if the technology was 99% accurate with white people but 95% accurate with people of color? What should happen? Should it not be deployed? Should white people be pulled aside the same proportion as people of color for human verification? Should it be deployed and priority be given to improving the accuracy in people of color?

Part of the problem with this logic is what you determine to be POC

Culturally, someone as white skinned as me could have grown up in Mexico in a Latino family with very Latino culture, and they would be a POC in everything but visuals. Meanwhile, my buddy is “Caucasian” and from San Francisco but his skin is dark as hell.

If you look at POC entirely from skin tone, facial shape, and hair, and no other metrics whatsoever, the “westernization” the US has on its citizens, as well as racial intermixing over 100+ years means that people that may very well be POC look white to a computer. So think of what the computer can figure out, I would argue probably 80% of the US is white to a computer, and the remaining 20% is a much smaller sample size so it is understandable that the computer is getting it wrong.

Reply on news.ycombinator.com