Amazon's "Rekognition" system. Even the spelling of the word is sinister

Martyn Warwick
By Martyn Warwick

Sep 30, 2019

via Flickr © zigazou76  (CC BY 2.0)

via Flickr © zigazou76 (CC BY 2.0)

  • Company is drafting its own facial recognition technology laws to present to Congress
  • Wants to "help" legislators in framing regulation
  • Self-interest and profit-protection masquerading as concern and altruism
  • Those to be governed by legislation should not write that legislation

In yet another example of the madness currently enveloping much of the world, Amazon, the manufacturer of a facial recognition system that invades privacy whilst being riddled with fundamental weaknesses, imperfections and has been proven to have inbuilt racist and sexist biases, is actually proposing that it should be allowed to draft facial recognition legislation and regulation and present it to the US Congress.

The US has a long and inglorious history of allowing corporations to have undue influence over the framing of laws that might affect their profitability and operations and it is just the same in this case. Amazon's facial recognition technology has so far found to be wanting so many times in so many ways that it makes perfect sense for its CEO, Jeff Bezos, to try to detract from those manifold failures by 'helping' to write legislation around the routine AI surveillance of masses of the population, even though the AI itself has been shown to be at fault up to 98 per cent of the time. Writing the laws yourself and in your own favour means scrutiny can be watered-down and policing and penalties can be minimised.

Last week, Mr. Bezos, a man who possesses a phizzog of such singularity that absolutely no AI is required to be able to recognise him, told an audience at an Amazon event in Seattle, Washington state that he has listened to demands by Congress that facial recognition technology must be regulated and has hand-picked a team of employees to draw up draft legislation to help the elected representatives of the people to come to the right conclusions.

He said. "Our public policy team is actually working on facial recognition regulations, and it makes a lot of sense to regulate that. It's a perfect example of something that has really positive uses, so you don't want to put the brakes on it. But at the same time, there's also potential for abuses of that kind of technology, so you do want regulations." As Raymond Chandler so succinctly put it, "No shit, Sherlock".

Oh, and by the way, millions of people want to put the brakes on facial recognition.

Reaction was immediate. The American Civil Liberties Union (ACLU) commented, "If Amazon is really interested in preventing these dangers, the first thing it should do is stop pushing surveillance tools into our communities without regard for the impact. Lawmakers should be skeptical of weak industry proposals that sacrifice individuals rights in the interest of profit."

Rekognition in Amerika. A totalitarian konstruct for diminished democracy?

Amazon isn't the only provider of facial recognition technology, there are several others including Microsoft, but Amazon's "Rekognition" drew a deluge of criticism last month when, in an ACLU test of the technology held in California, out of a database of 25,000 mugshots of wanted felons the system misidentified 20 per cent of the State's politicians as criminals on the run. Last year, in another test Rekognition flagged 28 members of the US Congress as wanted criminals.

And Amazon is not needed anyway. US legislators have already put forward several bills on the control of facial recognition systems, including the total banning of it in public places and on business premises. Indeed, Microsoft, showing a little more nous and awareness of how unpopular untrammelled use of the technology by organisations as diverse as the police, municipal authorities, shops and stores, car parking companies, parks and recreation management and private security companies, actually is, does at least admit that unless facial recognition laws are enacted the technology will soon "exacerbate societal issues." 

The question is, "should we trust corporations such as Amazon to draft its own laws and regulations when they stand to make billions of dollars in profits via a privacy-busting, all-encompassing surveillance-driven business model." The answer is, "No we should not." 

As Albert Cahn, the executive director of the US Surveillance Technology Oversight Project, says. "We can't trust the companies that have profited off of biased facial recognition systems for years to now write their own rules. Amazon's push for federal regulations is a cynical ploy to undercut the growing list of state laws that ban facial recognition like its own Rekognition system."

Time and time again Rekognition has been proven to be particularly biased against women and non-whites and very poor at actually identifying anyone. Nonetheless, the company continues to develop the technology. Recently, Amazon claimed that in addition to being able to identify faces according to gender, age, ethnicity, visible emotion and whether or not an individual is on some kind of a 'watch or detain' list, it can now detect fear itself. 

Soon we'll be being stopped by the cops because a facial recognition system reckons you are afraid of something, spiders perhaps, or flying, or even out-of-control and under-regulated facial recognition systems themselves. Mind you, after a few beers, a curry and a bowl of sauerkraut, the rictus of fear as analysed by Rekognition might actually be more to do with gas than abject terror. And that could easily be proven, right there in the street, and be a fine way to give facial recognition the blow off it so richly deserves.

Email Newsletters

Sign up to receive TelecomTV's top news and videos, plus exclusive subscriber-only content direct to your inbox.