iBorderCtrl is an AI based lie detector project funded by the European Union’s Horizon 2020. The tool will be used on people crossing borders of some European countries. It officially enables faster border control. It will be tested in Hungary, Greece and Letonia until August 2019 and should then be officially deployed.

The project will analyze facial micro-expressions to detect lies. We really have worries about such a project. For those who don’t have any knowledge on AI and CS, the idea of using a computer to detect lies can sound really good. Computers are believed to be totally objective.

But the AI community knows it is far from being true: biases are nearly omnipresent. We have no idea how the dataset used by iBorderCtrl has been built.

More globally, we have to remind that AI has no understanding of humans (to be honest, it has no understanding at all). It just starts being able to recognize the words we pronounce, but it doesn’t understand their meaning.

Lies rely on complex psychological mechanisms. Detecting them would require a lot more than a simple literal understanding. Trying to detect them using some key facial expressions looks utopian, especially as facial expressions can vary from a culture to another one. As an example, nodding the head usually means “yes” in western world, but it means “no” in countries such as Greece, Bulgaria and Turkey.

{ ActuIA | Continue reading }

The ‘iBorderCtrl’ AI system uses a variety of ‘at home’ pre-registration systems and real time ‘at the airport’ automatic deception detection systems. Some of the critical methods used in automated deception detection are that of micro-expressions. In this opinion article, we argue that considering the state of the psychological sciences current understanding of micro-expressions and their associations with deception, such in vivo testing is naïve and misinformed. We consider the lack of empirical research that supports the use of micro-expressions in the detection of deception and question the current understanding of the validity of specific cues to deception. With such unclear definitive and reliable cues to deception, we question the validity of using artificial intelligence that includes cues to deception, which have no current empirical support.

{ Security Journal | Continue reading }

The U.S. government is in the midst of forcing a standoff with China over the global deployment of Huawei’s 5G wireless networks around the world. […] This conflict is perhaps the clearest acknowledgement we’re likely to see that our own government knows how much control of communications networks really matters, and our inability to secure communications on these networks could really hurt us.

{ Cryptography Engineering | Continue reading }

related { Why Controlling 5G Could Mean Controlling the World }

When the world’s first cellular network was installed in Washington, D.C., in 1983, Gatt and Rwayitare knew instinctively that cellular could solve many of Africa’s communication problems. […] “How do you educate a government on what cellular is all about?” asked Gatt. […] You hand him a mobile phone and get him to call home, of course. […]

On an official state visit to the United States, Zairean President Mobutu Sese Seko was handed a Motorola mobile phone and urged to call home. It was 1985, so he had to be persuaded that the device — which weighed as much as a bottle of wine and boasted a retractable antenna — was not a walkie-talkie. But once he’d spoken to his family in Kinshasa, he needed no further convincing. Joseph Gatt and Miko Rwayitare’s plan was coming together.

Telecel, the company they formally founded a year later, would soon have 3,000 subscribers in Zaire (now the Democratic Republic of Congo) — well before mobile phones were ubiquitous in America. Zaire’s near-defunct fixed-line infrastructure meant that the country’s elite were willing to spend $5,000 on a handset and up to $16 per minute to remain connected. […]

Despite being impressed, the dictator — like most people in 1985 — hadn’t fully grasped how life-changing the technology would be and he initially refused to grant Telecel an operating license. Gatt and Rwayitare knew they were onto a good thing, however, so they used their life savings to purchase an ailing U.S. mobile technology firm and obtained finance from Motorola to erect a small system in Kinshasa. All that remained was to buy a couple of hundred handsets — at $3,000 a pop — and give them to Mobutu and his inner circle.

“These 200 Zairean officials called each other and overseas over the next year without paying for a single call.” […] At the end of the trial period, and faced with the prospect of losing what had now become an essential cog in the state machinery, Mobutu agreed to give them their license.

{ OZY | Continue reading }

Google has reportedly built a quantum computer more powerful than the world’s top supercomputers. A Google research paper was temporarily posted online this week, the Financial Times reported Friday, and said the quantum computer’s processor allowed a calculation to be performed in just over 3 minutes. That calculation would take 10,000 years on IBM’s Summit, the world’s most powerful commercial computer, Google reportedly said.

Google researchers are throwing around the term “quantum supremacy” as a result, the FT said, because their computer can solve tasks that can’t otherwise be solved. “To our knowledge, this experiment marks the first computation that can only be performed on a quantum processor,” the research paper reportedly said.

{ CNet | Continue reading }

photo { The Sphynx of Gizeh before excavation, photo taken by balloon, 1871 }

of course there is no behind the scenes, no real self, no authenticity, etc. just a precession of simulacra; influencers sort of serve the same function Baudrillard thought Disneyland served: to make everyone else feel “authentic”

{ Rob Horning }

We are connecting everything to everything.

[…]

In the network economy the winner-take-all behavior of Hollywood hit movies will become the norm for most products—even bulky manufactured items. Oil wells are financed this way now; a few big gushers pay for the many dry wells. You try a whole bunch of ideas with no foreknowledge of which ones will work. Your only certainty is that each idea will either soar or flop, with little in between. A few high-scoring hits have to pay for all the many flops. This lotterylike economic model is an anathema to industrialists, but that’s how network economies work. There is much to learn from long-term survivors in existing hits-oriented business (such as music and books). They know you need to keep trying lots of things and that you don’t try to predict the hits, because you can’t.

Two economists proved that hits—at least in show biz—were unpredictable. They plotted sales of first-run movies between May 1985 and January 1986 and discovered that “the only reliable predictor of a film’s box office was its performance the previous week. Nothing else seemed to matter—not the genre of the film, not its cast, not its budget.” The higher it was last week, the more likely it will be high this week— an increasing returns loop fed by word of mouth recommendations. The economists, Art De Vany and David Walls, claim these results mirror a heavy duty physics equation known as the Bose-Einstein distribution. The fact that the only variable that influenced the result was the result from the week before, means, they say, that “the film industry is a complex adaptive system poised between order and chaos.” In other words, it follows the logic of the net: increasing returns and persistent disequilibrium.

[…]

Because prices move inexorably toward the free, the best move in the network economy is to anticipate this cheapness.

{ Kevin Kelly, New Rules for the New Economy, 1998 | PDF | More: Wired }

Paul Hildreth peered at a display of dozens of images from security cameras surveying his Atlanta school district and settled on one showing a woman in a bright yellow shirt walking a hallway.

A mouse click instructed the artificial-intelligence-equipped system to find other images of the woman, and it immediately stitched them into a video narrative of her immediate location, where she had been and where she was going.

There was no threat, but Hildreth’s demonstration showed what’s possible with AI-powered cameras. If a gunman were in one of his schools, the cameras could quickly identify the shooter’s location and movements, allowing police to end the threat as soon as possible, said Hildreth, emergency operations coordinator for Fulton County Schools.

AI is transforming surveillance cameras from passive sentries into active observers that can identify people, suspicious behavior and guns, amassing large amounts of data that help them learn over time to recognize mannerisms, gait and dress. If the cameras have a previously captured image of someone who is banned from a building, the system can immediately alert officials if the person returns.

{ LA Times | Continue reading }

installation sketch { ecstasy, 2018 }

The team say that sextortion emails demanding cryptocurrency payment first appeared in 2018. The scammers send their emails via botnets, such as Necurs or Cutwail. These are global networks of computers infected with malware that send out spam on demand.

This is offered as a service on the dark net. Various researchers have shown that spammers pay botnet owners between $100 and $500 to send a million spam emails. They can even rent botnets at a cost of $10,000 per month, which allows them to send 100 million spam messages. […]

Back in 2008, one group of cyber-crime experts infiltrated a botnet for 26 days and monitored spammers sending 350 million emails for a pharmaceutical product. The result was 28 sales. This generated a revenue of $2,732, which corresponds to a conversion rate of just 0.00001%. Nevertheless, the experts concluded that by using additional botnets, spammers could generate around $9,500 per day which adds up to $3.5 million per year.

Sextortion has the potential to be much more profitable, say Paquet-Clouston and co. The reason is that it does not require the spammers to host any kind of e-commerce website, or to procure, store, and ship products of any kind. And cryptocurrency payments are simpler than bank payments and do not require the involvement of a friendly bank.

{ Technology Review | Continue reading }

oil on canvas { Caleb Brown, Sports Explosion, 2009 }

Tesla is a car company whose stock trades like a tech company. Tesla might sell 400,000 cars this year. By contrast, Ford might sell 6 million, GM 8.5 million. Granted, the Tesla Model 3 looks and drives like a dream. But when you count salaries and overhead according to Tesla’s own quarterly statements, it costs more to make a Tesla than people are willing to pay for it. And that calculus includes the federal subsidies that will dry up on December 31 of this year. Ford is worth $35 billion and makes money on its cars. Tesla is worth $40 billion and doesn’t. How is this math possible?

Tesla’s stock trades at such a large multiple of its revenue because Musk has convinced shareholders that it’s not a car company, but an artificial-intelligence company that happens to use a fleet of 500,000 cars to collect and label data. It’s a clever sleight-of-hand, but it’s not fooling those who matter. As a fund manager on Wall Street once told me, “You’re not a hedge-fund manager until you’ve shorted Tesla at least once.” […]

We estimate that ninety percent of the startups in the autonomous-vehicle space today will not exist in five years. […] The big crunch is coming because, over the next year, all the major auto and trucking companies will decide on who will be the suppliers for their main production lines in 2022. This won’t be for full self-driving, but for something a little more modest if still vitally important: a car so safe it is incapable of crashing.

{ National Review | Continue reading }

Before you hand over your number, ask yourself: Is it worth the risk? […]

Your phone number may have now become an even stronger identifier than your full name. I recently found this out firsthand when I asked Fyde, a mobile security firm in Palo Alto, Calif., to use my digits to demonstrate the potential risks of sharing a phone number.

He quickly plugged my cellphone number into a public records directory. Soon, he had a full dossier on me — including my name and birth date, my address, the property taxes I pay and the names of members of my family.

From there, it could have easily gotten worse. Mr. Tezisci could have used that information to try to answer security questions to break into my online accounts. Or he could have targeted my family and me with sophisticated phishing attacks.

{ NY Times | Continue reading }

image { Bell telephone magazine, March/April 1971 }

An artificial intelligence system should be recognised as the inventor of two ideas in patents filed on its behalf, a team of academics says.

The AI has designed interlocking food containers that are easy for robots to grasp and a warning light that flashes in a rhythm that is hard to ignore.

Patents offices insist innovations are attributed to humans - to avoid legal complications that would arise if corporate inventorship were recognised.

The academics say this is “outdated”.

{ BBC | Continue reading }

enamel on linen { Christopher Wool, Untitled, 2007 }

We analyzed over one million posts from over 4,000 individuals on several social media platforms, using computational models based on reward reinforcement learning theory. Our results consistently show that human behavior on social media qualitatively and quantitatively conforms to the principles of reward learning.

{ PsyArXiv | Continue reading }

image { Dissecting Reinforcement Learning }

A few months ago, I started hearing about something called Superhuman. It’s an invitation-only service that costs $30 a month and promises “the fastest email experience ever made.” Marc Andreessen, the influential venture capitalist, reportedly swore by it, as did tech bigwigs like Patrick and John Collison, the founders of Stripe. The app was rumored to have a waiting list of more than 100,000 people.

“We have the who’s who of Silicon Valley at this point,” Superhuman’s founder, Rahul Vohra, told me in an interview. The waiting list is actually 180,000 people long, he said, and some people are getting desperate. He showed me a photo of a gluten-free cake sent to Superhuman’s office by a person who was hoping to score an invitation.

“We have insane levels of virality that haven’t been seen since Dropbox or Slack,” Mr. Vohra added.

Last month, Superhuman raised a $33 million investment round, led by Mr. Andreessen’s firm, Andreessen Horowitz. That valued the company at roughly $260 million — a steep valuation for an app with fewer than 15,000 customers, but one apparently justified by the company’s trajectory and its support among fans, which borders on evangelical. […]

I spent several weeks testing it out. And it turns out that the hype is mostly justified, at least if you’re the kind of person who can spend $30 a month to get your inbox in order. […]

Some of the app’s features — such as ones that let users undo sending, track when their emails are opened and automatically pull up a contact’s LinkedIn profile — are available in other third-party email plug-ins. But there are bells and whistles that I hadn’t seen before. Like “instant intro,” which moves the sender of an introductory email to bcc, saving you from having to manually re-enter that person’s address. Or the scheduling feature, which sees that you’re typing “next Tuesday” and automatically pulls up your calendar for that day.

{ NY Times | Continue reading }

{ Horns are growing on young people’s skulls, caused by the forward tilt of the head. Phone use is to blame, research suggests. | Washington Post | No, Teenagers Are Not Growing ‘Skull Horns’ Because of Smartphones | Time }

Grow Your Own Cloud is a new service that helps you store your data nature’s way — in the DNA of plants.

We are at the forefront of the development of a new type of cloud, one that is organic, rather than silicon, and which emits oxygen rather than CO2.

{ GrowYourOwn.Cloud | Continue reading | Thanks Tim}

YouTube’s most lucrative channel in the UK did not feature a pop superstar, a legendary rock band or the biggest TV show around . . . but a child from Lancashire.

Gabriella — or Gaby, as she is known to her 12m-plus subscribers […] could be pulling in between £2,100 and £40,000 a day, depending on daily views, with earnings of up to £1m a year. […]

To novices like me, YouTube’s advertising system is incredibly complex and opaque. Fundamentally, the more popular you are, the more advertisers are prepared to pay. But you need to clock up 4,000 hours of viewing time for your videos in a year and have at least 1,000 subscribers to qualify for the Google Partner Programme, which provides a share of ad revenue.

Vloggers who choose cost-per-month advertising get a share ranging from 5p to £7 for every 1,000 “monetised” views — that is after YouTube takes 45 per cent. Only 40 per cent to 60 per cent of overall views are deemed worthy of advertising spending, based on various factors such as the location of the viewers and their level of “ad engagement”, that is, how much they skip or click on ads. […]

Competition on YouTube is brutal, with more than 450 hours of video uploaded every minute. Most of the experts who spoke to the Financial Times agreed that one or two videos a week was the bare minimum now required for success. Dan Middleton, a British gaming star who earned £25m last year, recently admitted working 12-hour days to post daily videos and stay ahead of the algorithm, which was recently tweaked to reward busier creators. Just missing a day can drastically reduce your views, subscribers and overall revenue.

Small wonder that “influencer burnout” is on the rise. […]

One vlogger with about 80,000 subscribers, who wished to remain anonymous, told me: “My most popular videos get six-figure views but the maximum I earn is a few hundred pounds. When I think about all the research, time, skills and promotion I put into each video, I wonder just how sustainable it can be.” […]

But YouTube advertising is not the only game in town. Creators increasingly make extra cash from influencer marketing. […]

Successful vloggers, and the agents managing their careers, can be cagey about the sums earned from influencer marketing. However, prying some numbers from leading agencies, I found that someone with at least 7m subscribers could earn about £250,000 for a sponsored video, while someone notching up 1m subscribers would command about £100,000. Even so-called “micro-influencers”, defined as having between 3,000 and 100,000 followers on any given platform, can earn between £2,000 and £50,000 for sponsored posts or videos.

{ Financial Times | Continue reading }

concrete with beehive structure, wax, and live bee colony { Pierre Huyghe, Untilled (Liegender Frauenakt), 2012 }

Behavioural patterns of Londoners going about their daily business are being tracked and recorded an unprecedented scale, internet expert Ben Green warns. […]

Large-scale London data-collection projects include on-street free Wi-Fi beamed from special kiosks, smart bins, police facial recognition and soon 5G transmitters embedded in lamp posts.

Transport for London announced this week they would track, collect and analyse movements of commuters around 260 Tube stations starting from July by using mobile Wi-Fi data and device MAC addresses to help improve journeys. Customers can opt out by turning off their Wi-Fi. 

{ Standard | Continue reading }

previously { The Business of Selling Your Location }

art { Poster for Autechre by the Designers Republic, 2016 }

Human-robot interaction in workplaces is a research area which remains unexplored.

In this paper, we present the results and analysis of a social experiment we conducted by introducing a humanoid robot (Nadine) into a collaborative social workplace.

The humanoid’s primary task was to function as a receptionist and provide general assistance to the customers. Moreover, the employees who interacted with Nadine were given over a month to get used to her capabilities, after which, the feedback was collected from the staff on the grounds of influence on productivity, affect experienced during interaction and their views on social robots assisting with regular tasks.

Our results show that the usage of social robots for assisting with normal day-to-day tasks is taken quite positively by the co-workers and that in the near future, more capable humanoid social robots can be used in workplaces for assisting with menial tasks.

{ PsyArXiv | Continue reading }

related { Is an Army of Robots Marching on Chinese Jobs? }

art { Hajime Sorayama }

In Shenzhen, the local subway operator is testing various advanced technologies backed by the ultra-fast 5G network, including facial-recognition ticketing.

At the Futian station, instead of presenting a ticket or scanning a QR bar code on their smartphones, commuters can scan their faces on a tablet-sized screen mounted on the entrance gate and have the fare automatically deducted from their linked accounts. […]

Consumers can already pay for fried chicken at KFC in China with its “Smile to Pay” facial recognition system, first introduced at an outlet in Hangzhou in January 2017. […]

Chinese cities are among the most digitally savvy and cashless in the world, with about 583 million people using their smartphones to make payment in China last year, according to the China Internet Network Information Center. Nearly 68 per cent of China’s internet users used a mobile wallet for their offline payments.

{ South China Morning Post | Continue reading }

photo { The Collection of the Australian National Maritime Museum }

Multidisciplinary studio curiosity has completed the flagship store of streetwear brand hipanda in omotesando, tokyo, combining digital and analogue features. The immersive retail interior brings together architectural elements with AR (augmented reality) and AE (augmented experience) technology in a sequence of spaces, inviting the visitor to look for the ‘host’ of the house, who is revealed through different interactive experiences, some digital and other analogue.

Curiosity has decked the hipanda store façade with the brand’s logo, which ‘jumps’ towards passersby with visual effects displayed through animations. inside, the main room features a play of light in constant motion, while the gallery space introduces products in a constant movement through AR, continually changing the presentation of the displays to bringing attention to the collection. The room’s walls are half-clad in mirrors, blurring the perception between digital and virtual.

{ designboom | Continue reading }

photo { Butcher shop specialized in game meat and poultry exhibiting a camel, Paris, 1908 }