Google

Google Chrome Proposes 'Privacy Sandbox' To Reform Advertising Evils 56

Google's Chrome team proposed a "privacy sandbox" Thursday that's designed to give us the best of both worlds: ads that publishers can target toward our interests but that don't infringe our privacy. From a report: It's a major development in an area where Chrome, the dominant browser, has lagged competitors. Browsers already include security sandboxes, restrictions designed to confine malware to limit its possible damage. Google's proposed privacy sandbox would similarly restrict tracking technology, according to proposal details Google published.

The privacy sandbox is "a secure environment for personalization that also protects user privacy," said Justin Schuh, a director of Chrome Engineering focused on security matters, in a privacy sandbox blog post. "Our goal is to create a set of standards that is more consistent with users' expectations of privacy." For example, Chrome would restrict some private data to the browser -- an approach rival Brave Software has taken with its privacy-focused rival web browser. And it could restrict sharing personal data until it's shared across a large group of people using technologies called differential privacy and federated learning.
Iphone

Chicago Tribune Claims iPhone Radiofrequency Radiation Levels Measured Higher Than Legal Safety Limit In Tests (macrumors.com) 74

An anonymous reader quotes a report from MacRumors: The Chicago Tribune recently launched an investigation into the radiofrequency radiation levels output by popular smartphones, and found that some of Apple's iPhones are allegedly emitting radiofrequency radiation that exceeds safety limits. According to the newspaper, it contracted an accredited lab to test several smartphones according to federal guidelines. iPhones were secured below clear liquid formulated to simulate human tissue while probes measured the radiofrequency radiation the liquid absorbed.

Several iPhones measured over the legal safety limits in the tests, but the worst performer was the iPhone 7. Its radiofrequency radiation exposure was over the legal limit and more than double what Apple reported to federal regulators. The iPhone X was slightly over limits in some tests, as was the iPhone 8, while the 8 Plus stayed within the legal range. iPhones were tested twice after Apple provided feedback on the testing method. The modified test "added steps intended to activate sensors designed to reduce the phones' power." In these modified tests, where a reporter held the iPhone to activate the sensors in question, the iPhone 8 was under the 5mm limit, but the iPhone 7 models were not. Apple disputed the results found by The Chicago Tribune and said that the lab did not test the iPhones in the same way that Apple does, though Apple would not specify what was done wrong in the testing. Apple also said the modified testing had been done wrong.
The investigation also found smartphones from Samsung, Motorola, and Vivo also demonstrated radiofrequency radiation levels that exceed FCC guidelines. However, it's worth noting that testing was done in a way to simulate the worst possible exposure conditions.

The FCC said that it will be conducting its own tests over the next couple of months. "We take seriously any claims on non-compliance with the RF (radiofrequency) exposure standards and will be obtaining and testing the subject phones for compliance with FCC rules," agency spokesman Neil Grace said.
AI

Amazon, Microsoft Are 'Putting World At Risk of Killer AI,' Says Study (ibtimes.com) 95

oxide7 shares a report from International Business Times: Amazon, Microsoft and Intel are among leading tech companies putting the world at risk through killer robot development, according to a report that surveyed major players from the sector about their stance on lethal autonomous weapons. Dutch NGO Pax ranked 50 companies by three criteria: whether they were developing technology that could be relevant to deadly AI, whether they were working on related military projects, and if they had committed to abstaining from contributing in the future.

Google, which last year published guiding principles eschewing AI for use in weapons systems, was among seven companies found to be engaging in "best practice" in the analysis that spanned 12 countries, as was Japan's Softbank, best known for its humanoid Pepper robot. Twenty-two companies were of "medium concern," while 21 fell into a "high concern" category, notably Amazon and Microsoft who are both bidding for a $10 billion Pentagon contract to provide the cloud infrastructure for the U.S. military. Others in the "high concern" group include Palantir, a company with roots in a CIA-backed venture capital organization that was awarded an $800 million contract to develop an AI system "that can help soldiers analyze a combat zone in real time." The report noted that Microsoft employees had also voiced their opposition to a U.S. Army contract for an augmented reality headset, HoloLens, that aims at "increasing lethality" on the battlefield.
Stuart Russel, a computer science professor at the University of California, argued it was essential to take the next step in the form of an international ban on lethal AI, that could be summarized as "machines that can decide to kill humans shall not be developed, deployed, or used."
Ruby

Backdoor Code Found In 11 Ruby Libraries (zdnet.com) 36

Maintainers of the RubyGems package repository have yanked 18 malicious versions of 11 Ruby libraries that contained a backdoor mechanism and were caught inserting code that launched hidden cryptocurrency mining operations inside other people's Ruby projects. ZDNet reports: The malicious code was first discovered yesterday inside four versions of rest-client, an extremely popular Ruby library. According to an analysis by Jan Dintel, a Dutch Ruby developer, the malicious code found in rest-client would collect and send the URL and environment variables of a compromised system to a remote server in Ukraine. "Depending on your set-up this can include credentials of services that you use e.g. database, payment service provider," Dintel said.

The code also contained a backdoor mechanism that allowed the attacker to send a cookie file back to a compromised project, and allow the attacker to execute malicious commands. A subsequent investigation by the RubyGems staff discovered that this mechanism was being abused to insert cryptocurrency mining code. RubyGems staff also uncovered similar code in 10 other projects. All the libraries, except rest-client, were created by taking another fully functional library, adding the malicious code, and then re-uploading it on RubyGems under a new name. All in all, all the 18 malicious library versions only managed to amass 3,584 downloads before being removed from RubyGems.

Businesses

Splunk To Buy Cloud-Monitoring Software Maker SignalFx For $1.05 Billion (wsj.com) 6

Splunk Inc. reached a $1.05 billion deal to buy cloud-monitoring startup SignalFx Inc., a deal that would strengthen the cybersecurity and data-analytics firm's offerings in the fast-growing cloud-computing sector. The Wall Street Journal reports: Founded in 2004, Splunk -- a play on the word "spelunking" -- collects and analyzes data to help companies identify patterns, like customers' beverage preferences, and detect anomalies, say fraud or a cyberattack. Splunk officials told analysts that Splunk has some customer overlap with San Mateo, Calif.-based SignalFx and that the target company's software represents a "top tier asset to the things that matter" to clients. Closely held SignalFx was valued at nearly $500 million after a $75 million funding round that closed in May, according to a Dow Jones VentureSource estimate.

The cash-and-stock deal is expected to close in the second half of Splunk's fiscal year, which ends Jan. 31. San Francisco-based Splunk, which went public in 2012 and carries a nearly $1.5 billion deficit, said it would be able to absorb the added operating costs from the deal. Splunk has been increasing its cloud business, which accounted for 25% of revenues in the July quarter and is expected to represent half of operations over the next few years, company officials said.

Communications

Google Postpones Shutdown of Hangouts For G Suite Users (theverge.com) 13

Google will let G Suite customers continue to use Hangouts until next year, delaying a shutdown of the service that was supposed to begin in October. Hangouts will now stay around for business customers until at least next June. The Verge reports: The shutdown will move customers of Google's business-focused G Suite subscription over to a pair of new chat services: Hangouts Chat, a Slack competitor; and Hangouts Meet, a video conferencing service. While the services generally include the same functionality (and more), people are pretty used to Hangouts, and Google says it's heard from companies' IT teams that they'd "like more time to migrate [their] organizations from classic Hangouts to Hangouts Chat."

Google says it now plans to start transitioning all G Suite users over to the new services by the end of next year. To make the transition easier, Google says it's going to work on adding more features to classic Hangouts. Right now, classic Hangouts users can only directly message a Hangouts Chat user. In the future, Google suggests that classic users may be able to view or participate in group chats, too.

Education

California High School In Silicon Valley Is Locking Up Students' Cellphones (nbcnews.com) 319

San Mateo High School administrators have instituted a new policy to lock up students' cellphones. "Each school day, nearly 1,700 students place their devices in a Yondr pouch that closes with a proprietary lock," reports NBC News. "School administrators unlock them at the end of the day." The goal is to help students focus on the teacher and other students. From the report: While administrators and teachers say they have already noticed a positive effect on students, the policy has elicited mixed reactions from researchers who argue its long-term effectiveness. Devices remain in the student's possession, but they aren't able to access them, the school said. The program was funded with a $20,000 grant. The pouches have been assigned to students at no cost, but losing one will cost the high-schoolers a $25 replacement fee.

Some technology experts feel the new policy is a step in the right direction and will curb distraction in the classroom. "Taking cellphones out of the classroom is a no-brainer," said Calvin Newport, a professor of computer science at Georgetown University. Students tend to perform worse when they have access to network connectivity in the classroom, he said. "The ability to be free of distraction and concentrate on things is increasingly valuable, so it's a good general function of our schools to be a place where our students get trained to keeping their concentration on one thing at a time," he added.

While many researchers have focused on the benefits of cutting out devices from the classroom, others worry about taking away something young people depend on. Larry Rosen, a research psychologist at California State University, said young people constantly check their phones to alleviate anxiety. They are anxious about staying on top of things, and that anxiety will build up if they are forced to ditch the devices cold turkey, he added. Taking away phones doesn't work for everyone, he argues. Instead, he believes "technology breaks" are a much happier medium.

Transportation

Waymo Releases a Self-Driving Open Data Set For Free Use By Research Community (techcrunch.com) 10

An anonymous reader quotes a report from TechCrunch: Waymo is opening up its significant stores of autonomous driving data with a new Open Data Set it's making available for the purposes of research. The data set isn't for commercial use, but its definition of "research" is fairly broad, and includes researchers at other companies as well as academics. The data set is "one of the largest, riches and most diverse self-driving data sets ever released for research," according to Waymo principal scientist and head of Research, Drago Anguelov, who was at both Zoox and Google prior to joining Waymo last year. Anguelov said in a briefing that the reason he initiated the push to make this data available is that Waymo and several other companies working in the field are "currently hampered by the lack of suitable data sets."

The Waymo Open Data set tries to fill in some of these gaps for their research peers by providing data collected from 1,000 driving segments done by its autonomous vehicles on roads, with each segment representing 20 seconds of continuous driving. It includes driving done in Phoenix, Ariz.; Kirkland, Wash.; Mountain View, Calif.; and San Francisco, Calif., and offering a range of different driving conditions, including at night, during rain, at dusk and more. The segments include data collected from five of Waymo's own proprietary lidars, as well as five standard cameras that face front and to the sides, providing a 360-degree view captured in high resolution, as well as synchronization Waymo uses to fuse lidar and imaging data. Objects, including vehicles, pedestrians, cyclists and signage is all labeled.
"We decided to contribute our part to make, ultimately, researchers in academia ask the right questions -- and for that, they need the right data," Anguelov said. "And I think this will help everyone in the field; it is not an admission in any way that we have problems solving these issues. But there is always room for improvement in terms of efficiency, scaleability, amount of labels to need. It's a developing field. Mostly we're trying to get others into thinking about our problems and working with us, as opposed to doing work that's potentially not so impactful, given the current state of things."
Businesses

DoorDash Still Pockets Workers' Tips Almost a Month After It Promised To Stop (vox.com) 81

DoorDash, the leading food delivery app in the U.S., is still pocketing workers' tips, despite announcing last month that it would stop the practice and change its tipping policies. The announcement was made after a report from The New York Times highlighted how the company uses tips to make up the worker's base pay -- essentially stealing the money you're trying to give someone to maximize their profits. Vox reports: At the time, CEO Tony Xu announced in a series of tweets that DoorDash would institute a new model to ensure workers' earnings would "increase by the exact amount a customer tips on every order." Xu promised to provide "specific details in the coming days." The next day, Xu sent out a note to DoorDash workers, broadly outlining changes and letting them know âoewhat to expect in the days ahead."

But 27 days later, current DoorDash workers tell Recode that the company's pay and tipping policies have stayed the same. The company has not made any public statements about its worker pay and how it plans to institute the changes, nor has it offered a specific date when it will fulfill its promise. A spokesperson declined to comment about the company's plans to change its tipping policy.
Soon after DoorDash's years-long tipping scheme was mentioned in the NYT's report, a class-action lawsuit was filed against the company for misleading its customers about how their tips were used. The lawsuit, filed at the end of July, claims that DoorDash failed to make clear to its customers that tips they gave through its app to couriers were not being allocated as they were intended to be, and that had customers known this, they would not have tipped through the app.
AI

Google DeepMind Co-Founder Placed On Leave From AI Lab (bloomberg.com) 14

Mustafa Suleyman, the co-founder of Google's high-profile AI lab DeepMind, has been placed on leave after controversy over some of the projects he led. Bloomberg reports: Mustafa Suleyman runs DeepMind's "applied" division, which seeks practical uses for the lab's research in health, energy and other fields. Suleyman is also a key public face for DeepMind, speaking to officials and at events about the promise of AI and the ethical guardrails needed to limit malicious use of the technology. "Mustafa is taking time out right now after 10 hectic years," a DeepMind spokeswoman said. She didn't say why he was put on leave.

He founded DeepMind in 2010 alongside current Chief Executive Officer Demis Hassabis. Four years later, Google bought DeepMind for 400 million pounds (currently $486 million), an ambitious bet on the potential of AI that set off an expensive race in Silicon Valley for specialists in the field. DeepMind soon began working on health-care research, eventually creating a division dedicated to the area. Suleyman, nicknamed "Moose" and whose mother was a nurse, led the development of the DeepMind Health team, building it into a 100-person unit.

The Courts

Flaws in Cellphone Evidence Prompt Review of 10,000 Verdicts in Denmark (nytimes.com) 36

The authorities in Denmark say they plan to review over 10,000 court verdicts because of errors in cellphone tracking data offered as evidence. From a report: The country's director of public prosecutions on Monday also ordered a two-month halt in prosecutors' use of cellphone data in criminal cases while the flaws and their potential consequences are investigated. "It's shaking our trust in the legal system," Justice Minister Nick Haekkerup said in a statement. The first error was found in an I.T. system that converts phone companies' raw data into evidence that the police and prosecutors can use to place a person at the scene of a crime. During the conversions, the system omitted some data, creating a less-detailed image of a cellphone's whereabouts. The error was fixed in March after the national police discovered it. In a second problem, some cellphone tracking data linked phones to the wrong cellphone towers, potentially connecting innocent people to crime scenes, said Jan Reckendorff, the director of public prosecutions.

"It's a very, very serious case," Mr. Reckendorff told Denmark's state broadcaster. "We cannot live with incorrect information sending people to prison." The authorities said that the problems stemmed partly from police I.T. systems and partly from the phone companies' systems, although a telecom industry representative said he could not understand how phone companies could have caused the errors. The national police determined that the flaws applied to 10,700 court cases dating to 2012, but it is unclear whether the faulty data was a decisive factor in any verdicts. The justice minister set up a steering group to track the extent of the legal problems they may have caused and to monitor the reviews of cases that may have been affected.

Security

Intel, Google, Microsoft, and Others Launch Confidential Computing Consortium for Data Security (venturebeat.com) 44

Major tech companies including Alibaba, Arm, Baidu, IBM, Intel, Google Cloud, Microsoft, and Red Hat today announced intent to form the Confidential Computing Consortium to improve security for data in use. From a report: Established by the Linux Foundation, the organization plans to bring together hardware vendors, developers, open source experts, and others to promote the use of confidential computing, advance common open source standards, and better protect data. "Confidential computing focuses on securing data in use. Current approaches to securing data often address data at rest (storage) and in transit (network), but encrypting data in use is possibly the most challenging step to providing a fully encrypted lifecycle for sensitive data," the Linux Foundation said today in a joint statement. "Confidential computing will enable encrypted data to be processed in memory without exposing it to the rest of the system and reduce exposure for sensitive data and provide greater control and transparency for users."

The consortium also said the group was formed because confidential computing will become more important as more enterprise organizations move between different compute environments like the public cloud, on-premises servers, or the edge. To get things started, companies made a series of open source project contributions including Intel Software Guard Extension (SGX), an SDK for code protection at the hardware layer.

Privacy

MoviePass Exposed Thousands of Unencrypted Customer Card Numbers (techcrunch.com) 14

New submitter sizzlinkitty writes: Movie ticket subscription service MoviePass has exposed tens of thousands of customer card numbers and personal credit cards because a critical server was not protected with a password. Mossab Hussein, a security researcher at Dubai-based cybersecurity firm SpiderSilk, found an exposed database on one of the company's many subdomains. The database was massive, containing 161 million records at the time of writing and growing in real time. Many of the records were normal computer-generated logging messages used to ensure the running of the service -- but many also included sensitive user information, such as MoviePass customer card numbers. These MoviePass customer cards are like normal debit cards: they're issued by Mastercard and store a cash balance, which users who sign up to the subscription service can use to pay to watch a catalog of movies.
Facebook

How Flat Earthers Nearly Derailed a Space Photo Book (nytimes.com) 167

An anonymous reader writes: A photographer trying to raise money for a self-published book of historical space artifacts had his Facebook ads repeatedly removed by Facebook because flat-Earthers and Moon hoax conspiracy theorists were offended. About 24 hours after the ads were approved, he got a notification telling him the ad had been removed. He resubmitted it. It was accepted -- and then removed again -- 15 or 20 times, he said. The explanation given: He had run "misleading ads that resulted in high negative feedback."

He understood that it was Facebook's algorithm that rejected the ads, not a person. Getting additional answers proved difficult, a common complaint with advertising on Facebook. The best clues he could find came in the comments under the ads, which he and his colleagues captured in screenshots before they were removed and in responses to other posts about the project: There were phrases such as "The original moon landing was faking" and "It's all a show," along with memes mocking space technology. Some comments were hard to gauge, with users insisting that the earth was flat but that they'd buy the book anyway.

Mr. Redgrove didn't entirely blame the commenters. If these were their beliefs, then of course they were going to be annoyed by the ads. But how these individuals had ended up with the power to derail his campaign perplexed him. "They don't really have their systems in place to protect people," Mr. Redgrove said of Facebook. Facebook said it could not immediately comment on what had gone wrong. On Thursday, after the publication of this article, a representative for the company said it had investigated the issue and had confirmed that, as Mr. Redgrove had said, all the ads were originally approved.

AI

Gmail in G Suite Now Uses AI For Inline Spelling and Grammar Suggestions (venturebeat.com) 25

Ever been stumped by spelling or sentence syntax while pecking out an email to colleagues? Fortunately for G Suite users, Google will soon introduce improved spelling and grammar correction tools in Gmail that offer corrections as you type. From a report: Starting August 20 for rapid release domains and September 12 for scheduled release domains across all G Suite editions, Google will begin applying AI to make real-time spell-check suggestions while detecting potential grammar issues. For some common spelling mistakes, it'll also add "as-you-type" autocorrection for improved accuracy. The inline grammar suggestions are a carryover from Google Docs, which gained them back in February 2019. Squiggly blue lines appear under erroneous phrases as you write them, and right-clicking on them accepts or dismisses the corrections. The Mountain View company says its engine can handle basic cases like "affect" versus "effect" and "there" versus "their," in addition to more complicated rules like how to use prepositions correctly or to pick the right verb tense.
Chrome

Apple, Google, and Mozilla Block Kazakhstan's HTTPS Intercepting Certificate (zdnet.com) 80

Apple, Google, and Mozilla have moved in to ban a root certificate the Kazakhstan government used in the past month to spy on its citizens' web traffic. From a report: Starting today, Chrome, Firefox, and Safari will show errors if any HTTPS web traffic is encrypted with the Kazakh government's root or leaf certificates. This coordinated action will ensure the safety of Kazakh users who were forced last month by their local Kazakh ISPs to install this certificate under the threat of not being allowed to use the internet otherwise. Kazakh ISPs forced their customers to install the government's root certificate after the Kazakh government issued a decree and said the measure was "aimed at enhancing the protection of citizens, government bodies and private companies from hacker attacks, Internet fraudsters and other types of cyber threats." But in reality, the Kazakh government abused this root certificate installed in millions of users browsers to intercept and decrypt HTTPS traffic users were making to 37 domains, such as such Facebook, Google, Twitter, Instagram, and YouTube.
Anime

Japan's Digital Pop Stars Blur Line Between Virtual and Reality (wsj.com) 56

An anonymous Slashdot reader shares a report about Japan's virtual YouTubers or VTubers that act as live performers, corporate PR officials and even surrogate children. From The Wall Street Journal: Ryosei Takehisa, 24 years old, doesn't have any children -- unless you count an animated character with elfin ears called Mikuriya Kuon. In live appearances on YouTube, the kimono-clad Kuon character, voiced by an actor hired by Mr. Takehisa, dispenses advice about the latest video games and plays rock-paper-scissors with her fans. The creator says he considers Kuon his "real daughter" even though she "resides within pixels." While others may compete for fame or page views, "for me, I'm totally satisfied just with the fact that she was born and is continuing to live life in good health," says Mr. Takehisa. Digital avatars with human traits have long carved out a role on social media, on Instagram in particular. Japan, as it often does, has taken the idea and run with it, with its virtual characters now estimated to number more than 3,000.

Technology allows Kuon and her peers to have more direct engagement with fans -- and sometimes a family-like relationship with their own creators. The characters, known as virtual YouTubers or VTubers because many are active on YouTube, sing and dance at live performances and answer questions on webcasts. VTubers are so embedded in Japanese culture that one of them serves as a face of the Japanese government's tourism campaign. Another presented earnings results for game-site operator Gree Inc. in August last year, informing investors that "we will aggressively invest in strengthening our three earnings pillars."
"VTubers are an evolution in Japan's long tradition of manga and anime, giving real-time interactivity to the sort of characters earlier depicted in comic books and on television screens," the report says. "The next step could be artificial intelligence to allow the VTubers to sing, dance and be mischievous without any backstage human help."

Sony is trying to further extend one of their latest pop sensations, a VTuber called Kaguya Luna, by building on its virtual-reality technology. "It has already staged concerts by Luna that fans view through a VR headset," reports The WSJ. "Next the company is looking into haptic technology -- which can convey vibrations and force -- to allow fans to get up close and personal with Luna."
Security

The First Lightning Security Key For iPhones Is Here, and It Works With USB-C, Too (theverge.com) 51

Yubico is releasing the $70 YubiKey 5Ci, the first security key that can plug into your iPhone's Lightning port or a USB-C port, and it's compatible with popular password vaults LastPass and 1Password out of the box. The Verge reports: That means you may not have to remember your password for your bank ever again -- just plug the YubiKey into your iPhone, use it to log into the 1Password app, and get that bank password. At launch, it'll support these well-known password managers and single sign-on tools: 1Password, Bitwarden, Dashlane, Idaptive, LastPass, and Okta. And when using the Brave browser for iOS, the YubiKey 5Ci can be used as an easier way to log into Twitter, GitHub, 1Password's web app, and a couple other services.

Notably, the 5Ci doesn't work with the newest iPad Pros at all, despite having a USB-C connector that fits. And you can't just plug the Lightning side of the 5Ci into an iPhone and expect it to work with any service that supports the FIDO authentication protocol -- our passwordless future isn't here just yet. Yubico tells The Verge that services have to individually add support for Lightning connector on the 5Ci into their apps.

Technology

'World's Oldest Webcam' To Be Switched Off (bbc.com) 41

The world's oldest continuously working webcam is being switched off after 25 years. The BBC reports: The Fogcam was set up in 1994 to watch how the weather changed on the San Francisco State University campus. has broadcast almost continuously since then barring regular maintenance and the occasional need for it to be re-sited to maintain its view. Its creators said it was being shut down because there were now no good places to put the webcam. Jeff Schwartz, who with Dan Wong set up the webcam, said it would go offline on 30 August. "We felt it was time to let it go," Mr Schwartz told the SFGate newspaper, adding that it was getting harder to find secure locations to put the camera so it had a good view of Holloway Avenue.

Mr Schwartz said he was inspired to set up the camera by what is believed to be the first-ever live webcam which was set up at Cambridge University in 1993 to watch a communal coffee pot.
Robotics

YouTube Removes Videos of Robots Fighting For 'Animal Cruelty' (independent.co.uk) 94

YouTuber and robot enthusiast Anthony Murney noticed YouTube has removed hundreds of videos showing robots battling other robots after claiming they are in breach of its rules surrounding animal cruelty. He's blaming a new algorithm introduced by YouTube to detect instances of animal abuse. The Independent reports: Several other channels dedicated to robot combat have also produced videos pointing out the issue in an effort to get YouTube to restore the content. Channels posting robot combat videos saw their content removed and received a notice from YouTube explaining that the videos were in breach of its community guidelines. Each notice cited the same section of these guidelines, which states: "Content that displays the deliberate infliction of animal suffering or the forcing of animals to fight is not allowed on YouTube." It goes on to state: "Examples include, but are not limited to, dog fighting and cock fighting."

Slashdot Top Deals