AI

AI Can't Protect Us From Deepfakes, Argues New Report (theverge.com) 36

A new report from Data and Society raises doubts about automated solutions to deceptively altered videos, including machine learning-altered videos called deepfakes. Authors Britt Paris and Joan Donovan argue that deepfakes, while new, are part of a long history of media manipulation -- one that requires both a social and a technical fix. Relying on AI could actually make things worse by concentrating more data and power in the hands of private corporations. The Verge reports: As Paris and Donovan see it, deepfakes are unlikely to be fixed by technology alone. "The relationship between media and truth has never been stable," the report reads. In the 1850s when judges began allowing photographic evidence in court, people mistrusted the new technology and preferred witness testimony and written records. By the 1990s, media companies were complicit in misrepresenting events by selectively editing out images from evening broadcasts. In the Gulf War, reporters constructed a conflict between evenly matched opponents by failing to show the starkly uneven death toll between U.S. and Iraqi forces. "These images were real images," the report says. "What was manipulative was how they were contextualized, interpreted, and broadcast around the clock on cable television."

Today, deepfakes have taken manipulation even further by allowing people to manipulate videos and images using machine learning, with results that are almost impossible to detect with the human eye. Now, the report says, "anyone with a public social media profile is fair game to be faked." Once the fakes exist, they can go viral on social media in a matter of seconds. [...] Paris worries AI-driven content filters and other technical fixes could cause real harm. "They make things better for some but could make things worse for others," she says. "Designing new technical models creates openings for companies to capture all sorts of images and create a repository of online life."

Businesses

Workers Accuse Kickstarter of Union-Busting In Federal Complaint (vice.com) 44

On Monday night, unionizing employees at Kickstarter filed a complaint with the National Labor Review Board (NLRB) for allegedly wrongfully terminating two employees. Both of the employees were on the Kickstarter United organizing campaign. Motherboard reports: Kickstarter told Motherboard that the workers, Clarissa Redwine and Taylor Moore, were fired over performance issues within the past two weeks. But employees at Kickstarter are accusing the company of "discharging employees" because "they joined or supported a labor organization and in order to discourage union activities," according to the NLRB complaint, which was first reported and obtained by Slate's April Glaser. A third employee and member of the Kickstarter United organizing committee, Travis Brace, was informed on Thursday that he would no longer be needed in his role.

In a September 12 email obtained by Motherboard, Aziz Hasan, the CEO of Kickstarter, wrote to employees, "There have been allegations that we are retaliating against union organizing. Those allegations are not true. No Kickstarter employee has been or ever will be fired for union organizing." Redwine says the company complained to her in recent months that she was not satisfactorily working with her managers. She claims that she was not given specific guidance on how she could improve. "Suddenly, after becoming a public union organizer, I started to get very strong negative feedback," Redwine told Motherboard. "After my best quarter at the company, I was told I was being put on a Performance Improvement Plan for slippery reasons like not building trust with my managers. I asked how progress would be tracked over and over and only received answers akin to 'just trust us.' I assume they never crafted the Performance Improvement Plan because they couldn't come up with anything concrete for me to improve." Redwine and Moore are asking for back pay and to be reinstated to their positions.
In response to the complaint, Kickstarter said: "We'll be providing the NLRB with information about these firings and supporting documentation."

Kickstarter told Motherboard that it "recently terminated two employees for performance reasons. A third was working on a service we shut down, so his role was eliminated, and there were no other positions here that would be a strong fit. That staff member will be transitioning out of the company. All three of these employees were members of the organizing committee, but this has nothing to do with their departures. (We have fired three other people who were not organizers since March.)"

"We expect all employees -- including union organizers -- to be able to perform in their role and set up their teams and colleagues for success. We use a range of approaches -- twice-a-year performance reviews, peer feedback, manager feedback, one-on-one coaching and, in some cases, mediation -- to ensure that employees have the support they need to meet those expectations. When someone has been through this process and we have sufficient evidence that they are not meeting expectations, we must unfortunately part ways with them," the company continued.
Businesses

Apple Is Trying To Trademark 'Slofie' (theverge.com) 43

On Friday, Apple applied for a U.S. Trademark on "Slofie," a made-up name for slow-motion selfies, a feature that's new to the iPhone 11 models. "The phones' front camera can now record video at 120 frames per second, which when slowed down, results in a crisp slow-motion effect," writes Jacob Kastrenakes for The Verge. "The results are neat, though I'm not convinced they'll turn into the Animoji-like phenomenon Apple may be hoping for." From the report: Apple is applying for a trademark on slofies in connection with "downloadable computer software for use in capturing and recording video." That means this trademark seems to be more about preventing other companies from making slofie-branded camera apps than it is about limiting popular usage of this totally made-up word. Apple has reason to want to prevent the creation of knock-off slofie apps, too, since slofies are meant to be exclusive to the new iPhones.

Despite the focus on apps, Apple doesn't actually offer a slofie app or a slofie mode on the new iPhones. The feature is just called "slo-mo" in Apple's camera app, and the company's current usage of slofie refers exclusively to the resulting videos, not the app or mode used to capture them. Apple seems to be hoping slofies will be a fun selling point for its new phones. The feature is mentioned across Apple's website, and Apple presented a slofie ad during the phones' launch event. It wouldn't be surprising to see a lot more airing in the coming weeks once the phones are out.

The Courts

Facebook Plans Launch of Its Own 'Supreme Court' For Handling Takedown Appeals (arstechnica.com) 27

An anonymous reader quotes a report from Ars Technica: Facebook, which has managed to transcend geographic borders to draw in a population equal to roughly a third of all human life on Earth, has made its final charter for a "Supreme Court" of Facebook public. The company pledges to launch this initiative by November of next year. The new Oversight Board will have five key powers, according to a charter (PDF) Facebook released yesterday. It can "request that Facebook provide information" it needs in a timely manner; it can make interpretations of Facebook standards and guidelines "in light of Facebook's articulated values"; and it can instruct the company to allow or remove content, to uphold or reverse a decision leading to content being permitted or removed, and to issue "prompt, written explanations of the board's decisions."

"If someone disagrees with a decision we've made, they can appeal to us first, and soon they will be able to further appeal this to the independent board," company CEO Mark Zuckerberg wrote in a letter (PDF). "As an independent organization, we hope it gives people confidence that their views will be heard and that Facebook doesn't have the ultimate power over their expression." The board will launch with at least 11 members and should eventually get up to 40. The entity will contract its services to Facebook. Participants will serve a maximum of three three-year terms each and will be paid for their time. Their decisions will "be made publicly available and archived in a database of case decisions," with details subject to certain data or privacy restrictions. Facebook can also contact the board for an "automatic and expedited review" in exceptional circumstances, "when content could result in urgent real world consequences," such as, for example, if a mass-murderer is livestreaming his crimes. The panel's decisions will be binding, Facebook added, and the company will implement its findings promptly, "unless implementation of a resolution could violate the law."

Encryption

The FBI Tried To Plant a Backdoor in an Encrypted Phone Network (vice.com) 29

The FBI tried to force the owner of an encrypted phone company to put a backdoor in his devices, Motherboard has learned. From the report: The company involved is Phantom Secure, a firm that sold privacy-focused BlackBerry phones and which ended up catering heavily to the criminal market, including members of the Sinaloa drug cartel, formerly run by JoaquÃn "El Chapo" Guzman. The news signals some of the tactics law enforcement may use as criminals continue to leverage encrypted communications for their own ends. It also comes as Canadian media reported that a former top official in the Royal Canadian Mounted Police (RCMP), who has been charged with leaking state secrets, offered to sell information to Vincent Ramos, Phantom's CEO.

"He was given the opportunity to do significantly less time if he identified users or built in/gave backdoor access," one source who knows Ramos personally and has spoken with him about the issue after his arrest told Motherboard. A backdoor is a general term for some form of technical measure that grants another party, in this case the FBI, surreptitious access to a computer system. What exactly the FBI was technically after is unclear, but the desire for a backdoor was likely to monitor Phantom's clients.

IBM

IBM's New 53-qubit Quantum Computer is Its Biggest Yet (cnet.com) 44

IBM's 14th quantum computer is its most powerful so far, a model with 53 of the qubits that form the fundamental data-processing element at the heart of the system. From a report: The system, available online to quantum computing customers in October, is a big step up from the last IBM Q machine with 20 qubits and should help advance the marriage of classical computers with the crazy realm of quantum physics. Quantum computing remains a highly experimental field, limited by the difficult physics of the ultra-small and by the need to keep the machines refrigerated to within a hair's breadth of absolute zero to keep outside disturbances from ruining any calculations.

But if engineers and scientists can continue the progress, quantum computers could help solve computing problems that are, in practice, impossible on today's classical computers. That includes things like simulating the complexities of real-world molecules used in medical drugs and materials science, optimizing financial investment performance, and delivering packages with a minimum of time and fuel.

Security

Crypto-mining Malware Saw New Life Over the Summer as Monero Value Tripled (zdnet.com) 10

Malware that mines cryptocurrency made a comeback over the summer, with an increased number of campaigns being discovered and documented by cyber-security firms. From a report: The primary reason for this sudden resurgence is the general revival of the cryptocurrency market, which saw trading prices recover after a spectacular crash in late 2018. Monero, the cryptocurrency of choice of most crypto-mining malware operations, was one of the many cryptocurrencies that were impacted by this market slump. The currency also referred to as XMR, has gone down from an exchange rate that orbited around $300 - $400 in late 2017 to a meager $40 - $50 at the end of 2018. But as the Monero trading price recovered throughout 2018, tripling its value from $38 at the start of the year, to nearly $115 over the summer, so have malware campaigns.

These are criminal operations during which hackers infect systems with malware that's specifically designed to secretly mine Monero behind the computer owner's back. Starting with the end of May, the number of reports detailing crypto-mining campaigns published by cyber-security firms has exploded, with a new report published each week, and sometimes new campaigns being uncovered on a daily basis.

AI

AI Learned To Use Tools After Nearly 500 Million Games of Hide and Seek (technologyreview.com) 71

In the early days of life on Earth, biological organisms were exceedingly simple. They were microscopic unicellular creatures with little to no ability to coordinate. Yet billions of years of evolution through competition and natural selection led to the complex life forms we have today -- as well as complex human intelligence. Researchers at OpenAI, the San-Francisco-based for-profit AI research lab, are now testing a hypothesis: if you could mimic that kind of competition in a virtual world, would it also give rise to much more sophisticated artificial intelligence? From a report: The experiment builds on two existing ideas in the field: multi-agent learning, the idea of placing multiple algorithms in competition or coordination to provoke emergent behaviors, and reinforcement learning, the specific machine-learning technique that learns to achieve a goal through trial and error. In a new paper released today, OpenAI has now revealed its initial results. Through playing a simple game of hide and seek hundreds of millions of times, two opposing teams of AI agents developed complex hiding and seeking strategies that involved tool use and collaboration. The research also offers insight into OpenAI's dominant research strategy: to dramatically scale existing AI techniques to see what properties emerge.
Facebook

Facebook Contractors Have Been Listening To 'Hey Portal' Clips 13

Facebook, which last month said it stopped using humans to review and transcribe users' voice messages, will resume that practice for some audio collected from its Portal video-calling device. From a report: Facebook "paused human review of audio" around August. Bloomberg reported at the time the company hired contractors to transcribe private voice messages sent via its Messenger app. In that case, users had not been alerted to the possibility that their communications might be subject to human review. It was also unclear at the time that some of the clips Facebook had been collecting were coming from Portal.

Facebook confirmed Wednesday that it was indeed collecting audio from Portal users who make a request from the device using the command "Hey Portal." By default, those commands were recorded and stored on Facebook servers, and some of them were transcribed by contractors working with the company to improve the software algorithms used to understand the commands, according to Andrew Bosworth, Facebook's head of hardware. That practice was paused last month at the same time Messenger stopped using humans to transcribe messages.
Facebook

Facebook Expands Definition of Terrorist Organizations To Limit Extremism (nytimes.com) 145

Facebook on Tuesday announced a series of changes to limit hate speech and extremism on the social network, expanding its definition of terrorist organizations and planning to deploy artificial intelligence to better spot and block live videos of shooters. The company is also expanding a program that redirects users searching for extremism to resources intended to help them leave hate groups behind. The New York Times reports: The announcement came the day before a hearing on Capitol Hill on how Facebook, Google and Twitter handle violent content. Lawmakers are expected to ask executives how they are handling posts from extremists. In its announcement post, Facebook said the Christchurch tragedy "strongly" influenced its updates. And the company said it had recently developed an industry plan with Microsoft, Twitter, Google and Amazon to address how technology is used to spread terrorist accounts.

Facebook said that it had mostly focused on identifying organizations like separatists, Islamist militants and white supremacists. The company said that it would now consider all people and organizations that proclaim or are engaged in violence leading to real-world harm. The team leading its efforts to counter extremism on its platform has grown to 350 people, Facebook said, and includes experts in law enforcement, national security, counterterrorism and academics studying radicalization. To detect more content relating to real-world harm, Facebook said it was updating its artificial intelligence to better catch first-person shooting videos. The company said it was working with American and British law enforcement officials to obtain camera footage from their firearms training programs to help its A.I. learn what real, first-person violent events look like.

AT&T

AT&T Sued For Allegedly Inflating DirecTV Now Subscriber Numbers (engadget.com) 19

A lawsuit seeking class action status says AT&T is inflating AT&T TV Now -- previously known as DirecTV Now -- subscriber numbers by creating fake users. From a report: It's accusing the company's management of carrying out the scheme in an effort to make the service look good in the eyes of investors even though it was struggling with serious technical and financial problems. The management did so, according to the lawsuit, by encouraging employees to add DirecTV Now subscription fees to subscribers' accounts without their knowledge or consent. One of the methods employees allegedly used is tacking on up to three accounts to a single customer's phone number -- including those who just signed up for a free trial -- and running their credit card three times. In some cases, customers were reportedly charged for a subscription even though they made it clear that they didn't want it. Sales employees allegedly made and used fake email accounts in both instances.
Medicine

Millions of Americans' Medical Images and Data Are Available On the Internet (arstechnica.com) 22

An anonymous reader quotes a report from ProPublica: Medical images and health data belonging to millions of Americans, including X-rays, MRIs, and CT scans, are sitting unprotected on the Internet and available to anyone with basic computer expertise. The records cover more than 5 million patients in the United States and millions more around the world. In some cases, a snoop could use free software programs -- or just a typical Web browser -- to view the images and private data, an investigation by ProPublica and the German broadcaster Bayerischer Rundfunk found.

We identified 187 servers -- computers that are used to store and retrieve medical data -- in the U.S. that were unprotected by passwords or basic security precautions. The computer systems, from Florida to California, are used in doctors' offices, medical-imaging centers, and mobile X-ray services. The insecure servers we uncovered add to a growing list of medical records systems that have been compromised in recent years. Unlike some of the more infamous recent security breaches, in which hackers circumvented a company's cyber defenses, these records were often stored on servers that lacked the security precautions that long ago became standard for businesses and government agencies.
The exposed data varied depending on the health provider and the software they use. "For instance, the server of U.S. company MobilexUSA displayed the names of more than a million patients -- all by typing in a simple data query," reports ProPublica. "Their dates of birth, doctors, and procedures were also included."

"Another imaging system, tied to a physician in Los Angeles, allowed anyone on the Internet to see his patients' echocardiograms," the report adds. "All told, medical data from more than 16 million scans worldwide was available online, including names, birthdates, and, in some cases, Social Security numbers."

The authors of the report recommend you ask your health care provider or doctor if access to your images requires a login and password, and to ask if they conduct a regular security assessment as required by HIPAA.
Transportation

Modified Tesla Model S Reportedly Outguns Porsche Taycan at Nurburgring (cnet.com) 145

Tesla is being a busy bee at the Nurburgring Nordschleife, and after multiple videos and rumors, we have our first idea of how capable its modified Model S electric sedan is. From a report: Road & Track reported Tuesday that the stripped-down Model S clocked a lap time of 7:23. Do note, that is not an official time and comes from well-placed sources hand-timing the electric car with a clear view of the track. It also happens to be 20 seconds quicker than the Porsche Taycan's lap time at the 'Ring. However, what's impressive is that this time comes during an industry-pool session.

This is when all automakers are free to make rounds around the track, which leads to traffic. At the end of the day, the time doesn't reflect the car's full potential. It's not possible to clock a flying lap during an industry-pool session. If we do get an official time this week, expect it to be even quicker than 7:23. At the same time, the magazine's sources also described the fact that this Model S is nowhere near close to the kind of electric sedan Tesla sells to buyers. It's rumored to house a new three-motor powertrain (one motor for each rear wheel) and sports plenty of variations outside. They include wider tires, reportedly Goodyear Eagle F1 Supersport RS rubber, fender flares to help cover the extra width and a massive gurney flap at the rear. Massive carbon-ceramic brakes are supposedly hiding behind the wheels, too.

Facebook

Facebook Enlists Ray-Ban Maker To Help Develop Secret 'Orion' Smart Glasses (cnbc.com) 31

Facebook is partnering with Ray-Ban parent company Luxottica to help speed up development of its top-secret augmented reality glasses that are being developed out of its Facebook Reality Labs in Redmond, Washington. The social media company is hoping the new partnership will help them complete the glasses between 2023 and 2025. CNBC reports: The glasses are internally codenamed Orion, and they are designed to replace smartphones, the people said. The glasses would allow users to take calls, show information to users in a small display and live-stream their vantage point to their social media friends and followers. Facebook is also developing an artificial intelligence voice assistant that would serve as a user input for the glasses, CNBC previously reported. In addition, the company has experimented with a ring device that would allow users to input information via motion sensor. That device is code-named Agios.

The company has hundreds of employees at its Redmond offices working on technology for the AR glasses, but thus far, Facebook has struggled to reduce the size of the device into a form factor that consumers will find appealing, a person who worked on the device told CNBC. Given the long lead time, there's no guarantee that the glasses will be completed on time or ever ship. But one person familiar with the project said that CEO Mark Zuckerberg has a strong interest in the glasses, and asked hardware chief Andrew Bosworth to prioritize them.

Operating Systems

Fedora Linux 31 Beta is Here (betanews.com) 19

An anonymous reader shares a report: Fedora 31 is due later this year, but first, there needs to be some beta testing. And so, today, Fedora 31 Beta is made available for download. Unfortunately, details surrounding version 31 are a bit sparse. With that said, one big change involves Fedora users with ARM 64-based single board computers, such as a Raspberry Pi. Those folks will get access to an additional desktop spin -- the lightweight Xfce. Workstation users will be treated to GNOME 3.34, but not the final version that was released recently. Don't worry -- when Fedora leaves beta status, and is officially released, you can be sure the stable GNOME 3.34 will be included. Remember, Fedora is one of the best ways to experience a vanilla GNOME desktop environment.
Firefox

Firefox Moving To a Faster 4-Week Release Cycle (mozilla.org) 50

Mozilla announces in a blog post: We typically ship a major Firefox browser (Desktop and Android) release every 6 to 8 weeks. Building and releasing a browser is complicated and involves many players. To optimize the process, and make it more reliable for all users, over the years we've developed a phased release strategy that includes 'pre-release' channels: Firefox Nightly, Beta, and Developer Edition. With this approach, we can test and stabilize new features before delivering them to the majority of Firefox users via general release.

And today we're excited to announce that we're moving to a four-week release cycle! We're adjusting our cadence to increase our agility, and bring you new features more quickly. In recent quarters, we've had many requests to take features to market sooner. Feature teams are increasingly working in sprints that align better with shorter release cycles. Considering these factors, it is time we changed our release cadence. Starting Q1 2020, we plan to ship a major Firefox release every 4 weeks. Firefox ESR release cadence (Extended Support Release for the enterprise) will remain the same. In the years to come, we anticipate a major ESR release every 12 months with 3 months support overlap between new ESR and end-of-life of previous ESR. The next two major ESR releases will be ~June 2020 and ~June 2021.

Google

Amazon, Google, Facebook, and Microsoft Workers Have Pledged To Strike on September 20 (vice.com) 102

In a historic first, more than 1,200 Amazon employees announced last week that they will walk out of their offices on September 20 over the company's carbon footprint. In the following days, tech workers at Microsoft, Facebook, and Google have announced they will also go on strike. From a report: Of the big five tech companies, this leaves Apple employees, who have yet to make any statement on the climate strikes, which will take place in 120 countries. On the day its competitors are striking, Apple will release the iPhone 11. While Apple claims that the iPhone is getting greener, it's still a product that necessarily requires exploiting the Earth, and, in some cases, labor.
The Internet

The Internet Relies on People Working for Free (medium.com) 89

Who should be responsible for maintaining and troubleshooting open-source projects? From a report: When you buy a product like Philips Hue's smart lights or an iPhone, you probably assume the people who wrote their code are being paid. While that's true for those who directly author a product's software, virtually every tech company also relies on thousands of bits of free code, made available through "open-source" projects on sites like GitHub and GitLab. Often these developers are happy to work for free. Writing open-source software allows them to sharpen their skills, gain perspectives from the community, or simply help the industry by making innovations available at no cost. According to Google, which maintains hundreds of open-source projects, open source "enables and encourages collaboration and the development of technology, solving real-world problems."

But when software used by millions of people is maintained by a community of people, or a single person, all on a volunteer basis, sometimes things can go horribly wrong. The catastrophic Heartbleed bug of 2014, which compromised the security of hundreds of millions of sites, was caused by a problem in an open-source library called OpenSSL, which relied on a single full-time developer not making a mistake as they updated and changed that code, used by millions. Other times, developers grow bored and abandon their projects, which can be breached while they aren't paying attention. It's hard to demand that programmers who are working for free troubleshoot problems or continue to maintain software that they've lost interest in for whatever reason -- though some companies certainly try. Not adequately maintaining these projects, on the other hand, makes the entire tech ecosystem weaker. So some open-source programmers are asking companies to pay, not for their code, but for their support services. Daniel Stenberg is one of those programmers. He created cURL, one of the world's most popular open-source projects.

Businesses

Square Wants To Be a Bank But Doesn't Want To Be Taxed Like One (wsj.com) 111

An anonymous reader quotes a report from The Wall Street Journal: Square has lent $5 billion to small businesses and consumers and applied for a banking license. It operates a nationwide mobile money-transfer business that serves 15 million Americans. But Square says it isn't a financial company. The San Francisco-based payments processor filed a lawsuit last week against its home city to recover $1.3 million in taxes, plus interest and attorneys' fees. Square argued that San Francisco was wrong to classify it as a financial company for tax purposes because it is a technology company that should be subject to a lower tax rate. The refund Square is seeking from San Francisco relates to excess taxes it believes it paid for 2014 and 2015, years when Square was a much smaller company. Square disclosed in its most recent securities filing that it may have to pay San Francisco additional taxes for subsequent years if it was unable to convince the city it was a tech company.

The company said in its complaint that its San Francisco operations more closely resemble a technology company than a financial-services firm. The company has roughly 2,200 employees in San Francisco. Square also said in its lawsuit that San Francisco overstates Square's gross receipts because it includes money that the company doesn't get to keep. Square takes a 2.75% cut of most credit-card payments to its small-business customers. More than half of what Square derives from those transactions must be paid out to banks and credit-card networks, the company said.
San Francisco's treasurer and tax collector, Jose Cisneros, said that his office's audit of Square was "thorough and fair" and that he stands by its findings. Mr. Cisneros is also a defendant in the case.
Google

Google Preps 'Smart Screenshots' Feature To Let You Search With a Screenshot (9to5google.com) 17

According to Abner Li from 9to5Google, Google is working on a new "Smart Screenshots" feature that integrates Google Lens abilities into the Google app's screenshot function. From the report: The Google app has long had an "Edit & share screenshots" ability where captures made within Search would reveal cropping and annotation tools. Meanwhile, Assistant has long maintained a "What's on my screen" capability that analyzes what you're currently viewing for search suggestions. Google app 10.61 reveals work on "Smart Screenshots" that combine those two features. Like before, a toolbar -- which interestingly uses a four-color light bar -- appears after you take a screenshot. A small preview is shown at the left with a pencil button overlaid. You can open the system share sheet, but the Google app also suggests a frequently used app.

The most interesting addition is Lens. "Exploring with Lens" could be intended as a "Screen search" replacement given that Lens is increasingly taking over visual lookup throughout first-party apps, like Chrome. After taking a capture, Smart Screenshots have an easy way to invoke Lens for search, OCR, and finding visually "similar items." The existing editing tools (Annotating, Cropping, and Sharing) will remain and this new functionality appears to even use the same settings toggle to enable. It's unclear if this functionality once live will again be limited to screenshots taken within Search, or if it will expand to be systemwide and invokable anywhere. A notification from the Google app could appear after capturing a screenshot.

Slashdot Top Deals