Connect with us

Tech News

With reward comes risk: facial recognition and privacy

(TECH NEWS) Facial recognition and artificial intelligence are awesome rewards from technical innovation but with reward comes risk.

Published

on

ai facial recognition risk

Technology is an omnipresent force in all of our lives. It is the core of innovation, providing us with quick, new ways to research, socialize and entertain ourselves. It seems like everyone is taking advantage of rapidly changing technology.

However what one person thinks as a reward of new systems may actually be a risk to someone else.

Take for instance, facial recognition software. Facebook uses it to identify familiar faces in photos and Apple uses it to unlock phones. It’s everywhere.

Even the porn industry is getting in on it. PornHub, a major online source for adult content, announced their new plan to use AI to help categorize the 10,000 plus videos that are uploaded every day.

Prior to this update, the site used a system of tagging videos to keep them organized. I would go into examples of such categories, but I’ll leave that up to the imagination.

One non-explicit example is organizing content based on the names of the stars of the film. Both the site itself and users had the ability to add tags to videos.

Regardless, this was not fast enough. By integrating AI software, PornHub hopes to expedite this process.

While this may sound like a smart business decision, this seems like high risk beginning to inadvertently diminish privacy rights.

Many people in the porn industry have alternate personas to separate their work and personal lives. Facial recognition software may pull from sources from both sides of that spectrum and end up merging the two.

This has already been the case on Facebook via the recommendations the site makes for “people you may know” via your internet practices.

However, it’s not just a matter of protecting the identity for a professional or amateur porn actor, it’s also about the privacy of clients.

Imagine being recommended to friend the star of the last video you streamed. This industry in particular, requires a level of discretion.

To combat some of the fears, PornHub has insists that the AI software only tags from the 10,000 stars in their database. Though as this update has proven, they could expand their database to keep up with the demand in the future.

It’s a technological advantage for their organization, but at what cost to others’ privacy?

Natalie is a Staff Writer at The American Genius and co-founded an Austin creative magazine called Almost Real Things. When she is not writing, she spends her time making art, teaching painting classes and confusing people. In addition to pursuing a writing career, Natalie plans on getting her MFA to become a Professor of Fine Art.

Tech News

Another thing that can trick iPhone X facial recognition

(TECH NEWS) The iPhone X has had an array of challenges, even with their innovative facial recognition technology.

Published

on

iphone X facial recognition

Yiiiikes, a mask tricked Apple’s new Face ID feature. Vietnamese security firm Bkav Corporation recently held a demo pointing out flaws in the iPhone X’s facial recognition, claiming the technology is not as secure as Apple originally touted.

Bkav Corporation focuses on network security, anti-virus software, and mobile security software. Bkav Corp created a 3D mask that “beat” Face ID in a demonstration. The mask was crafted with a combination of 3D printing and 2D images.

When verifying users, Face ID takes photos using infrared cameras. The first photo creates the surface of the face then the second pic makes a mesh, reproducing the face in 3D. From there, Face ID uses AI technology to distinguish faces.

The 2D/3D hybrid is meant to throw off the AI feature specifically. According to Bkav’s VP of Cyber Security Ngo Tuan Anh, “Apple’s AI can only distinguish either a 100% real face or a 100% fake one. So if you create a ‘half-real half-fake’ face, it can fool Apple’s AI.”

Face ID is supposed to have a one million-to-one chance of false recognition.

Compared to Touch ID’s potential fail rate of fifty thousand-to-one, Face ID is meant to be way more secure. However, the risk of a false recognition increases with identical twins, siblings, and children under the age of thirteen since their facial features aren’t finished developing.

When iPhone X launched, Apple stated they worked with professional Hollywood mask makers and makeup artists to ensure Face ID couldn’t be fooled by masks or other prosthetics. While Apple noted Face ID should still work if users get haircuts, change facial hair, or sometimes wear glasses, masks weren’t part of the good-to-go features for unlocking phones.

If you’re one of the adopters of iPhone X, don’t start freaking out yet though. To create their mask, Bkav had to use a handheld scanner to get pictures of their target’s face. As in, the person whose phone they were trying to hack had to be in the same room to get the initial scans.

Plus, Bkav could have intentionally done a subpar job of setting up the Face ID. The obvious solution if you’re still worried? Add a passcode as well and don’t trust anyone who wants to make a mask of your face.

Read also: Do literally anything with your money besides buy the iPhone X

Continue Reading

Tech News

Well great, now the robots can do acrobatics

(TECH NEWS) Do you want Terminators? Because this is how you get Terminators. Bipedal robots can do backflips now…

Published

on

robots y'all

This is it. It’s happening. Robots.

A year ago Boston Dynamics robot Atlas was learning to stand and falling over while walking. Now, Atlas has been upgraded, allowing it to easily scale blocks parkour style, doing backflips, and even raising its arms triumphantly after nailing a landing.

And I am raising a card with a 10 for the solid execution, albeit shakily, because the first thing that went through my head watching those eerily fluid, human-like movements, was imagining it stomping over piles of human skulls with an AK in its cold, calculating hands.

We can build it. We have the technology.

Let’s hope it doesn’t find videos all those videos on YouTube of its creators tormenting the thing; prodding it with hockey sticks like a lion tamer with a chair, knocking boxes out of its arms, pushing it over, and kicking its robo-dog companions.

Atlas won’t forget that.

Imagine this thing chasing you in the woods, or down the street. In a few years, I wouldn’t be surprised if the Atlas bots were wearing badges. Atlas is far more spry than the dopey droids you might find in a Star Wars flick, and well on its way to creating Skynet from Boston Dynamics.

Guys, Atlas can do acrobatics now, like a ninja:

Anywhere human feet can tread, an advanced enough droid will be able to go (can we start calling them droids now?). If you knock them over, they get right back up. Those human-powered mechs have nothing on Atlas. Give it enough time, and Atlas will run circles around both Eagle Prime and KURATAS. They won’t need us puny humans for robot battles.

One day, they might not need us at all.

All jokes aside, it’s an incredible, awe-inducing advancement in robotics. Boston Dynamics also recently revealed a smaller, less creepy version of their robo-dog Spot to bring us SpotMini: a small four-legged robot that can climb stairs and moves similarly to the way a dog would romp about.

Just so you know, this is nothing to be afraid of… We’ve only just found out that robots are evolving at an alarming rate.

Continue Reading

Tech News

Social media giants are trying to operate without human controls but are failing

(TECH NEWS) Artificial intelligence (AI) is taking over in fascinating ways, but this big experiment of replacing human tasks is failing. Good news / bad news.

Published

on

ai robot not human

Let me tell you a story. In fact, let me tell you several.

A village in Macedonia had a small economic boom during the 2016 election, plagiarizing and stitching together pro-Trump messages on social media, then publishing the results as “news” in order to profit from Google ad revenue.

Back during the “Keep Calm and…” T-shirt fad, a shirt company went through a thoroughly justified PR apocalypse for selling products labeled “Keep Calm and Hit Her” and “Keep Calm and Rape a Lot.”

The 17th most popular website on Earth occasionally likes to tell women over 30 to freeze their ova.

So! That’s a parade of fail. What’s it got in common, beyond making any reasonable reader consider moving to an Amish community and trying to forget even the word “Internet”?

People. More accurately, their absence.

Veles, Macedonia churned out profitable nonsense about Trump slapping a protester (that didn’t happen) or getting the blessing of the Pope (Pope says nope) because Google ads are programmatic. There’s no QA component, no human eyes reviewing content and asking “is this worth advertising on?” or for that matter “is this blatantly false?”

Likewise the Evil T-Shirt Crisis. The company generated slogans by dropping memes into an algorithm, then throwing the result on Amazon. That ended… poorly.

We, and every other tech and business network in the digital cosmos, have written in depth about all the dang robots taking our jobs. Usually our primary concern is the economic fallout. We’re a business news organization. It’s our job to warn you about that stuff.

But there’s another problem, and it’s a huge problem, especially as media consumption in general continues to rise, and more and more of that media is moderated by algorithms rather than people.

Robots aren’t just taking our jobs. They suck at our jobs. Algorithms may play go, but they aren’t ready to make value judgments yet. A quick Google will yield a dozen more examples of AI failures just as repulsive and/or hilarious as the ones on my list. And the real punchline for all of that?

It’s good news.

For once, the robot apocalypse is cutting us puny humans a break. It’s creating jobs almost as fast as it’s gobbling them up, because at this point, it is excruciatingly clear that robots aren’t ready to produce work people can actually see. They’re not even ready to put ads on work people see, not without causing a PR catastrophe every other month.

AI isn’t a better widget. It also isn’t an employee that doesn’t want benefits or take long lunches.

It’s a product in permanent beta, desperately trying to catch up to the constantly changing nuance of human interaction. It doesn’t work without homo sapiens holding its robot hand.

Let’s call it Salter’s Law: For every application of AI to customer-facing work, you will need to hire at least one human for damage control when the AI screws up.

Continue Reading
Advertisement

The
American Genius
News neatly in your inbox

Join thousands of AG fans and SUBSCRIBE to get business and tech news updates, breaking stories, and MORE!

Emerging Stories