Slashdot Mirror

An anonymous reader quotes a report from The Verge: Now that Google+ is history, today, Google unveiled what will be offered to G Suite users in its place: Currents. The new app "enables people to have meaningful discussions and interactions across your organization, helping keep everyone in the know and giving leaders the opportunity to connect with their employees." The company says Currents has a new look and feel compared to Google+ -- it seems somewhat similar to my eyes -- and it's been streamlined to make it faster to post content and tag it. Posts from a company's top executives can be given priority in the Currents stream to make sure employees see it. Currents is launching in beta, and Google says G Suite companies can request access to the program starting now. Google+ posts will automatically be transferred over to Currents. (I'm just talking about G Suite posts; personal Google+ posts are a goner at this point.) If the Currents name sounds familiar, it's because it "was previously a magazine app that was the precursor to Google Play Newsstand, which itself was later replaced by Google News," the report notes.

An anonymous reader quotes a report from Ars Technica: Google is releasing the second Android Q Beta today. As we learned with the first release, Android Q is bringing support for foldable smartphones, better privacy and permissions controls, and a grab bag of other features. We've yet to install the second beta on one of our own devices, but Google's release blog post promises "bug fixes, optimizations, and API updates," as well as a crazy new multitasking feature and an emulator for foldables. Android loves multitasking. So far we've had split screens and floating windows, and Android Q Beta 1 even had a hidden desktop mode. Beta 2 brings us a new multitasking feature called "Bubbles." Bubbles let you minimize an app into a little circle, which floats around on the screen above all your other apps. Tapping on a bubble will open a small UI. The only demo Google shows is one for a messaging app. Each bubble is a contact, and tapping on the bubble shows a small chat UI. If you remember Facebook's "Chat Head" UI for Messenger, Bubbles is that, but built into the OS. "Bubbles are great for messaging because they let users keep important conversations within easy reach," Google said in their blog post. "They also provide a convenient view over ongoing tasks and updates, like phone calls or arrival times. They can provide quick access to portable UI, like notes or translations, and can be visual reminders of tasks too."

Google's latest speech recognition works entirely offline, eliminating the delay that many other voice assistants have to return your query. "The delay occurs because your voice, or some data derived from it anyway, has to travel from your phone to the servers of whoever operates the service, where it is analyzed and sent back a short time later," reports TechCrunch. "This can take anywhere from a handful of milliseconds to multiple entire seconds (what a nightmare!), or longer if your packets get lost in the ether." The only major downside with Google's new system is its limited availability. As of right now, it's only available to people with a Pixel smartphone. From the report: Why not just do the voice recognition on the device? There's nothing these companies would like more, but turning voice into text on the order of milliseconds takes quite a bit of computing power. It's not just about hearing a sound and writing a word -- understanding what someone is saying word by word involves a whole lot of context about language and intention. Your phone could do it, for sure, but it wouldn't be much faster than sending it off to the cloud, and it would eat up your battery. But steady advancements in the field have made it plausible to do so, and Google's latest product makes it available to anyone with a Pixel.

Google's work on the topic, documented in a paper here, built on previous advances to create a model small and efficient enough to fit on a phone (it's 80 megabytes, if you're curious), but capable of hearing and transcribing speech as you say it. No need to wait until you've finished a sentence to think whether you meant "their" or "there" -- it figures it out on the fly. So what's the catch? Well, it only works in Gboard, Google's keyboard app, and it only works on Pixels, and it only works in American English. So in a way this is just kind of a stress test for the real thing. "Given the trends in the industry, with the convergence of specialized hardware and algorithmic improvements, we are hopeful that the techniques presented here can soon be adopted in more languages and across broader domains of application," writes Google in their blog post.

On Thursday, Google began officially selling their new .dev domains to anyone, Engadget reports: To claim a .dev, all you need to do is sign up with your registrar of choice (Google, naturally, is an option). As a bonus, Google is offering a free .dev domain to anyone who applied for a ticket to the Google I/O event happening this May.

The domains will be secure by default, as they all require HTTPS, and Google has already moved many of its own sites (including web.dev, opensource.dev and flutter.dev) to the domain.
"The internet has come a long way from the days of .com, .org and .net," writes Engadget. "Now, you can get domains ending in anything from .cool to .ninja."

UPDATE (3/3/2019): Apparently the domain emacs.dev now points to the web site for Vim.

An anonymous reader quotes a report from Bloomberg: Disney is said to have pulled its advertising spending from YouTube, joining other companies including Nestle, after a blogger detailed how comments on Google's video site were being used to facilitate a "soft-core pedophilia ring." Some of the videos involved ran next to ads placed by Disney and Nestle. All Nestle companies in the U.S. have paused advertising on YouTube, a spokeswoman for the company said Wednesday in an email. Video game maker Epic Games and German packaged food giant Dr. August Oetker KG also said they had postponed YouTube spending after their ads were shown to play before the videos. Disney has also withheld its spending.

On Sunday, Matt Watson, a video blogger, posted a 20-minute clip detailing how comments on YouTube were used to identify certain videos in which young girls were in activities that could be construed as sexually suggestive, such as posing in front of a mirror and doing gymnastics. Watson's video demonstrated how, if users clicked on one of the videos, YouTube's algorithms recommended similar ones. By Wednesday, Watson's video had been viewed more than 1.7 million times. Total ad spending on the videos mentioned was less than $8,000 within the last 60 days, and YouTube plans refunds, the spokeswoman said. Two years ago, Verizon, AT&T, Johnson & Johnson and other major companies pulled their ads from YouTube after learning that some of their ads surfaced next to extremist and violent content. Yesterday, YouTube released an updated policy about how it will handle content that "crosses the line" of appropriateness.

"Any content -- including comments -- that endangers minors is abhorrent and we have clear policies prohibiting this on YouTube. We took immediate action by deleting accounts and channels, reporting illegal activity to authorities and disabling violative comments," a spokeswoman for YouTube said in an email.

stoborrobots writes: Google Fiber is leaving Louisville, as reported in The Verge: "Google Fiber's attempt to roll out its gigabit internet across the city of Louisville, Kentucky has apparently failed so spectacularly that the company has decided to completely shut down the service and leave town altogether. CNET has a report on the news, which Alphabet's Access division confirmed in a blog post on Thursday. 'We'll work with our customers and partners to minimize disruption, and we're committed to doing right by the community, which welcomed us as we tested methods of delivering high-speed internet in new and different ways,' the Fiber team said."

TechCrunch's take: "It's a rare admission of defeat for Google Fiber, though it's no secret that the company isn't exactly bullish on the prospect of the service anymore. Louisville was supposed to be somewhat of a comeback for Google Fiber, which like so many Google services is now under more pressure to generate a profit. Clearly, that didn't work out." The issue apparently has to do with "shallow trenching," a process that involves laying fiber cable two inches beneath the sides of roads in the city and covering them up with sealant. "The company seemed optimistic about this plan until some of the cable started becoming exposed over time, requiring a second cover-up with hot asphalt," reports The Verge. "It seems Access realized it had to go a bit deeper with the cabling; in San Antonio, a similar method is used -- but the fiber is laid at least six inches deep into the ground."

"Unfortunately, things have somehow gone so awry in Louisville that Google Fiber claims it would need to rebuild the entire network to get everything to a satisfactory point, and it seems Alphabet just isn't interested in blowing the cash that would be necessary to do that. So instead, Google Fiber will today alert Louisville customers that their service will end on April 15th." In an attempt to soften the blow, Google Fiber says it will not charge customers for their final two months of service.

In a blog post, Google clarified the timeline of the transition from classic Hangouts to Chat and Meet for its paying G Suite customers. "For them, the Hangouts retirement party will start in October of this year," reports TechCrunch. From the report: For consumers, the situation remains unclear, but Google says there will be free versions of Chat and Meet that will become available "following the transition of G Suite customers." As of now, there is no timeline, so for all we know, Hangouts will remain up and running into 2020. As for G Suite users, Google says it will start bringing more features from classic Hangouts to Chat between April and September. Those include integration with Gmail, the ability to talk to external users, improved video calling and making calls with Google Voice.

Most Android smartphone vendors have been notorious for the time they take to roll out the newest Android OS updates to their respective handsets. To tackle this, Google in 2017 announced Project Treble, which bypasses some middlemen in delivering new updates to consumers. With Project Treble now supported by all Android phone makers, in theory updates should roll out to us faster than before. To test this, news blog AndroidAuthority looked at the data to see where things stand. From the report: On average, Nougat updates took about 192 days to reach key devices, while Oreo was slightly faster at 170. Android Pie updates hit devices much faster, averaging just 118 days from Google's launch to significant OEM rollout. That's a significant improvement, though we're still waiting on updates from LG and HTC, which could drag this average back up. Most manufacturers are faster at providing updates now, but a few are slower. Huawei, Samsung, and Xiaomi were noticeably quicker this time around, bringing updates to key devices before the end of 2018. OnePlus and Sony were especially fast, but they've always been speedier than most. Disappointingly, Motorola has rolled out updates to its flagship Z series slower over the last few years.

An anonymous reader quotes a report from Ars Technica: Google is implementing major new Play Store rules for how Android's "SMS" and "Call Log" permissions are used. New Play Store rules will only allow certain types of apps to request phone call logs and SMS permissions, and any apps that don't fit into Google's predetermined use cases will be removed from the Play Store. The policy was first announced in October, and the policy kicks in and the ban hammer starts falling on non-compliant apps this week.

Google says the decision to police these permissions was made to protect user privacy. SMS and phone permissions can give an app access to a user's contacts and everyone they've ever called, in addition to allowing the app to contact premium phone numbers that can charge money directly to the user's cellular bill. Despite the power of these permissions, a surprising number of apps ask for SMS or phone access because they have other, more benign use cases. So to clean up the Play Store, Google's current plan seems to be to (1) build more limited, replacement APIs for these benign use cases that don't offer access to so much user data and (2) kick everyone off the Play Store who is still using the wide-ranging SMS and phone permissions for these more limited use cases. Google provides a help page that helps explain the new rules and offer workarounds for some use cases.

An anonymous reader quotes a report from Ars Technica: Google is implementing major new Play Store rules for how Android's "SMS" and "Call Log" permissions are used. New Play Store rules will only allow certain types of apps to request phone call logs and SMS permissions, and any apps that don't fit into Google's predetermined use cases will be removed from the Play Store. The policy was first announced in October, and the policy kicks in and the ban hammer starts falling on non-compliant apps this week.

Google says the decision to police these permissions was made to protect user privacy. SMS and phone permissions can give an app access to a user's contacts and everyone they've ever called, in addition to allowing the app to contact premium phone numbers that can charge money directly to the user's cellular bill. Despite the power of these permissions, a surprising number of apps ask for SMS or phone access because they have other, more benign use cases. So to clean up the Play Store, Google's current plan seems to be to (1) build more limited, replacement APIs for these benign use cases that don't offer access to so much user data and (2) kick everyone off the Play Store who is still using the wide-ranging SMS and phone permissions for these more limited use cases. Google provides a help page that helps explain the new rules and offer workarounds for some use cases.

Android Studio 3.3 is now available to download through stable channel, Google said Monday. The top new features of Android Studio 3.3 include a navigation editor, profiler tracking options, improvements on the build system, and lazy task configuration. However, the big focus with the new version was on "refinement and quality," the company said. Further reading: VentureBeat.

Google is pulling support for Android 4.0 Ice Cream Sandwich more than seven years after it was first introduced. The company announced in a blog post that Google Play services will no longer provide updates for the APIs (14 and 15) used by applications running on ICS. VentureBeat reports: Ice Cream Sandwich (ICS), as Android 4.0 to 4.0.4 is more affectionately known, was a landmark operating system in many ways, ushering in a whole new set of interface guidelines -- with a more minimalist design, not to mention groundbreaking features such as near-field communication (NFC), lockscreen support for camera and music controls, and facial recognition smarts for unlocking devices.

App developers who currently offer minimum support of API level 16 (Android 4.1 Jelly Bean) and over won't have to do anything as a result of these changes. However, if their apps currently support API level 14 or 15, they will encounter a build error when updating to a newer SDK version. Google is now recommending that all developers target API level 16 as the bare minimum, which means those still using Ice Cream Sandwich on their Android device won't even see the app update in Google Play, let alone be able to download it.

YouTube CEO Susan Wojcicki has again hit out at proposed new European Union copyright rules which she claims is impossible for a platform like YouTube to comply with, and if done so, could harm the creative industries. Wojcicki said the European Parliament's vote in favor of an overhaul to copyright law two months ago is "unrealistic" because owners often disagree on who owns the rights to online material. In a blog post, she wrote: Take the global music hit "Despacito." This video contains multiple copyrights, ranging from sound recording to publishing rights. Although YouTube has agreements with multiple entities to license and pay for the video, some of the rights holders remain unknown. That uncertainty means we might have to block videos like this to avoid liability under article 13. Multiply that risk with the scale of YouTube, where more than 400 hours of video are uploaded every minute, and the potential liabilities could be so large that no company could take on such a financial risk.

The consequences of article 13 go beyond financial losses. EU residents are at risk of being cut off from videos that, in just the last month, they viewed more than 90bn times. Those videos come from around the world, including more than 35m EU channels, and they include language classes and science tutorials as well as music videos. We welcome the chance to work with policymakers and the industry to develop a solution within article 13 that protects rights holders while also allowing the creative economy to thrive. This could include more comprehensive licensing agreements, collaboration with rights holders to identify who owns what, and smart rights management technology, similar to Content ID.

A new investigation uncovers a sophisticated ad fraud scheme involving more than 125 Android apps and websites, some of which were targeted at kids. From a report: Last April, Steven Schoen received an email from someone named Natalie Andrea who said she worked for a company called We Purchase Apps. She wanted to buy his Android app, Emoji Switcher. But right away, something seemed off. "I did a little bit of digging because I was a little sketched out because I couldn't really find even that the company existed," Schoen told BuzzFeed News. The We Purchase Apps website listed a location in New York, but the address appeared to be a residence. "And their phone number was British. It was just all over the place," Schoen said. It was all a bit weird, but nothing indicated he was about to see his app end up in the hands of an organization responsible for potentially hundreds of millions of dollars in ad fraud, and which has funneled money to a cabal of shell companies and people scattered across Israel, Serbia, Germany, Bulgaria, Malta, and elsewhere.

Schoen had a Skype call with Andrea and her colleague, who said his name was Zac Ezra, but whose full name is Tzachi Ezrati. They agreed on a price and to pay Schoen up front in bitcoin. "I would say it was more than I had expected," Schoen said of the price. That helped convince him to sell. A similar scenario played out for five other app developers who told BuzzFeed News they sold their apps to We Purchase Apps or directly to Ezrati. (Ezrati told BuzzFeed News he was only hired to buy apps and had no idea what happened to them after they were acquired.) The Google Play store pages for these apps were soon changed to list four different companies as their developers, with addresses in Bulgaria, Cyprus, and Russia, giving the appearance that the apps now had different owners.

But an investigation by BuzzFeed News reveals that these seemingly separate apps and companies are today part of a massive, sophisticated digital advertising fraud scheme involving more than 125 Android apps and websites connected to a network of front and shell companies in Cyprus, Malta, British Virgin Islands, Croatia, Bulgaria, and elsewhere. More than a dozen of the affected apps are targeted at kids or teens, and a person involved in the scheme estimates it has stolen hundreds of millions of dollars from brands whose ads were shown to bots instead of actual humans. (A full list of the apps, the websites, and their associated companies connected to the scheme can be found in this spreadsheet.)

One way the fraudsters find apps for their scheme is to acquire legitimate apps through We Purchase Apps and transfer them to shell companies. They then capture the behavior of the app's human users and program a vast network of bots to mimic it, according to analysis from Protected Media, a cybersecurity and fraud detection firm that analyzed the apps and websites at BuzzFeed News' request. This means a significant portion of the millions of Android phone owners who downloaded these apps were secretly tracked as they scrolled and clicked inside the application. By copying actual user behavior in the apps, the fraudsters were able to generate fake traffic that bypassed major fraud detection systems. Response from Google.

YouTube is creating a new Learning Fund program where it will invest $20 million toward education content. The announcement was made today by Malik Ducard, global head of learning. The Verge reports: Channels like TED-Ed, dedicated to educational Ted Talks, and Hank and John Green's Crash Course have already secured additional funding, according to YouTube's blog post. The company plans to invest in content from independent creators, like the Green brothers, as well as traditional news sources and educational organizations to broaden its content offering.

YouTube's Learning Fund has a nice ring to it, but it isn't a philanthropic charity. An FAQ about the program states that "successful applicants must enter into a written agreement with YouTube. This agreement will contain more details about required deliverables, payment timelines, and other terms and conditions." Creators must maintain a minimum of 25,000 subscribers. Those applying to the program also don't need to have a degree or proper certification in their field, "but successful applicants will be required to demonstrate that they have expertise and/or that the content they produce is verified by an expert in the field."

Google has announced a new feature for Google Assistant: it's bilingual now. The digital assistant will automatically recognize what language is being spoken to it and respond appropriately -- all without requiring you to change any settings. The Verge reports: You'll be able to set up Assistant to understand and respond to any two of the following languages: English, German, French, Spanish, Italian, and Japanese. Google adds that it intends on "expanding to more languages in the coming months." The behind-the-scenes tech to make this happen is pretty interesting, as Google explains in an accompanying blog post. To make Assistant receptive to two languages simultaneously, the company created a new language-identification model (which it calls LangID) that runs as soon as the software detects speech.

Assistant actually runs LangID in parallel with two separate language processing models that try to transcribe what's been said in the user's two preset languages. Once LangID has identified the language, Assistant then cancels the incorrect transcription and routes all processing power to focus on the correct one. In order to speed up the process of identification, LangID doesn't just consider vocabulary; it also signals the frequency at which each language is used and the type of device it's used with.

An anonymous reader writes: "Google has patched a vulnerability in the Chrome browser that allows an attacker to retrieve sensitive information from other sites via audio or video HTML tags," reports Bleeping Computer. The attack breaks CORS -- Cross-Origin Resource Sharing, a browser security feature that prevents sites from loading resources from other websites -- and will attempt to load resources (some of which can reveal information about users) inside audio and video HTML tags. During tests, a researcher retrieved age and gender information from Facebook users, but another researcher says the bug can be also used to retrieve data from corporate backends or private APIs. Ron Masas, a security researcher with Imperva, first discovered and reported this issue to Google. The bug was fixed at the end of July with the release of Chrome v68.0.3440.75.

An anonymous reader quotes PCWorld: The critical Meltdown and Spectre bugs baked deep into modern computer processors will have ramifications on the entire industry for years to come, and Chrome just became collateral damage. Google 67 enabled "Site Isolation" Spectre protection for most users, and the browser now uses 10 to 13 percent more RAM due to how the fix behaves.

"Site Isolation does cause Chrome to create more renderer processes, which comes with performance tradeoffs," Googleâ(TM)s Charlie Reis says. "On the plus side, each renderer process is smaller, shorter-lived, and has less contention internally, but there is about a 10-13% total memory overhead in real workloads due to the larger number of processes. Our team continues to work hard to optimize this behavior to keep Chrome both fast and secure." It's a significant performance hit, especially for a browser battling a reputation for being a memory hog, but a worthwhile one nonetheless.
Chrome's Spectre-blocking site isolation "is now enabled by default for 99 percent of Chrome users on all platforms."

In an effort to reduce misinformation on YouTube, the video-sharing website will be adding "authoritative" context to search results about conspiracy-prone topics, as well as putting $25 million toward news outlets producing videos. YouTube made the announcement today as part of a new step in its Google News Initiative, a journalism-focused program that aims to help publishers earn revenue and combat fake news. The Verge reports: This update includes new features for breaking news updates and long-standing conspiracy theories. YouTube is implementing a change it announced in March, annotating conspiracy-related pages with text from "trusted sources like Wikipedia and Encyclopedia Britannica." And in the hours after a major news event, YouTube will supplement search results with links to news articles, reasoning that rigorous outlets often publish text before producing video.

YouTube is also funding a number of partnerships. It's establishing a working group that will provide input on how it handles news, and it's providing money for "sustainable" video operations across 20 markets across the world, in addition to expanding an internal support team for publishers.

Brian Krebs reports: Google is reminding organizations to review how much of their Google Groups mailing lists should be public and indexed by Google.com. The notice was prompted in part by a review that KrebsOnSecurity undertook with several researchers who've been busy cataloging thousands of companies that are using public Google Groups lists to manage customer support and in some cases sensitive internal communications. Google Groups is a service from Google that provides discussion groups for people sharing common interests. Because of the organic way Google Groups tend to grow as more people are added to projects -- and perhaps given the ability to create public accounts on otherwise private groups -- a number of organizations with household names are leaking sensitive data in their message lists. Many Google Groups leak emails that should probably not be public but are nevertheless searchable on Google, including personal information such as passwords and financial data, and in many cases comprehensive lists of company employee names, addresses and emails. Google has outlined instructions on how to secure the discussion boards.

An anonymous reader writes: Tech blogger John Gruber appears to have successfully identified one of the restaurants mentioned in a post on Google's AI blog that bragged about "a meal booked through a call from Duplex." Mashable then asked a restaurant employee there if Google had let him know in advance that they'd be receiving a call from their non-human personal assistant AI. "No, of course no," he replied. And "When I asked him to confirm one more time that Duplex had called...he appeared to get nervous and immediately said he needed to go. He then hung up the phone."

John Gruber now asks: "How many real-world businesses has Google Duplex been calling and not identifying itself as an AI, leaving people to think they're actually speaking to another human...? And if 'Victor' is correct that Hong's Gourmet had no advance knowledge of the call, Google may have violated California law by recording the call." Friday he added that "This wouldn't send anyone to prison, but it would be a bit of an embarrassment, and would reinforce the notion that Google has a cavalier stance on privacy (and adhering to privacy laws)."
The Mercury News also reports that legal experts "raised questions about how Google's possible need to record Duplex's phone conversations to improve its artificial intelligence may come in conflict with California's strict two-party consent law, where all parties involved in a private phone conversation need to agree to being recorded."

For another perspective, Gizmodo's senior reviews editor reminds readers that "pretty much all tech demos are fake as hell." Speaking of Google's controversial Duplex demo, she writes that "If it didn't happen, if it is all a lie, well then I'll be totally disappointed. But I can't say I'll be surprised."

Christina Chiou Yeh, writing for Google Registry: On May 1 we announced .app, the newest top-level domain (TLD) from Google Registry. It's now open for general registration so you can register your desired .app name right now. We begin our journey with sitata.app, which provides real-time travel information about events like protests or transit strikes. Looks all clear, so our first stop is the Caribbean, where we use thelocal.app and start exploring. After getting some sun, we fly to the Netherlands, where we're feeling hungry. Luckily, picnic.app delivers groceries, right to our hotel. With our bellies full, it's time to head to India, where we use myra.app to order the medicine, hygiene, and baby products that we forgot to pack. Did we mention this was a business trip? Good thing lola.app helped make such a complex trip stress free. Time to head home now, so we slip on a hoodie we bought on ov.app and enjoy the ride.

At its I/O developer conference today, Google debuted "Duplex," an AI system for accomplishing real world tasks over the phone. "To show off its capabilities, CEO Sundar Pichai played two recordings of Google Assistant running Duplex, scheduling a hair appointment and a dinner reservation," reports Quartz. "In each, the person picking up the phone didn't seem to realize they were talking to a computer. The conversations proceed back-and-forth to find the right time, and confirm what the customer wanted. Even when conversations didn't go as expected, the assistant understood the context, responded appropriately, and carried on the task. (You can listen to the recordings here.)" From the report: It's a far more natural conversation than consumers may be used to with digital assistants. The AI's voice lacks a stilted cadence and comes complete with "ums" and natural pauses (which also helps cover up the fact that it is still processing). It uses the phone's on-board processing, as well as the cloud, to deliver the right response with just the right amount of pause.

Google is taking advantage of its primary asset: data. It trained Duplex on a massive body of "anonymized phone conversations," according to a release. Every scheduling task will have its own problems to solve when arranging a specific type of appointment, but all will be underpinned by Google's massive volume of data from searches and recordings that will help the AI hold a conversation. Still, the technology cannot carry on just any conversation. Even though Duplex can seemingly handle far more context than other systems, it only works within a narrow set of queries (Google hasn't listed all of them yet). And despite releasing six new more natural sounding voices for the Assistant product available today, none approached the humanity of its Duplex example.

At its I/O developer conference today, Google debuted "Duplex," an AI system for accomplishing real world tasks over the phone. "To show off its capabilities, CEO Sundar Pichai played two recordings of Google Assistant running Duplex, scheduling a hair appointment and a dinner reservation," reports Quartz. "In each, the person picking up the phone didn't seem to realize they were talking to a computer. The conversations proceed back-and-forth to find the right time, and confirm what the customer wanted. Even when conversations didn't go as expected, the assistant understood the context, responded appropriately, and carried on the task. (You can listen to the recordings here.)" From the report: It's a far more natural conversation than consumers may be used to with digital assistants. The AI's voice lacks a stilted cadence and comes complete with "ums" and natural pauses (which also helps cover up the fact that it is still processing). It uses the phone's on-board processing, as well as the cloud, to deliver the right response with just the right amount of pause.

Google is taking advantage of its primary asset: data. It trained Duplex on a massive body of "anonymized phone conversations," according to a release. Every scheduling task will have its own problems to solve when arranging a specific type of appointment, but all will be underpinned by Google's massive volume of data from searches and recordings that will help the AI hold a conversation. Still, the technology cannot carry on just any conversation. Even though Duplex can seemingly handle far more context than other systems, it only works within a narrow set of queries (Google hasn't listed all of them yet). And despite releasing six new more natural sounding voices for the Assistant product available today, none approached the humanity of its Duplex example.

First Google banned ads from payday lenders in 2016, now it will no longer allow ads from bail-bond companies. Ars Technica reports: In a blog post, the company suggested that such ads constitute a "deceptive or harmful product," citing a 2016 study concluding that minority and low-income communities are typically most affected by such services. "For-profit bail-bond providers make most of their revenue from communities of color and low-income neighborhoods when they are at their most vulnerable, including through opaque financing offers that can keep people in debt for months or years," Google wrote. Also in 2016, another study found that "there are 646,000 people locked up in more than 3,000 local jails throughout the U.S.," simply for their inability to pay a bond, which is what drives many people to the services of a bondsman. The change will take effect in July 2018.

An anonymous reader quotes a report from UploadVR: Google Research this week revealed an AR microscope (ARM) capable of detecting cancerous cells in real-time with the help of machine learning. Locating cancer with a standard microscope is a difficult and time-consuming process, with a raft of information for doctors to study and investigate. With this new solution, though, the microscope is able to quickly locate cancerous cells and then highlight them as a doctor peers inside. The platform uses a modified light microscope integrated with image analysis and machine learning algorithms into its field of view. An AR display sits above a camera that communicates with the algorithm to display data as soon as it locates an issue. In order words, the microscope immediately begins looking for cancerous cells as soon as you place a sample beneath it. It's effectively doing the same job as a doctor just, according to Google, a lot faster. Google posted a video about the AR microscope on YouTube.

An anonymous reader quotes a report from VentureBeat: Google today launched Chrome 66 for Windows, Mac, Linux, and Android. The desktop release includes autoplaying content muted by default, security improvements, and new developer features. You can update to the latest version now using the browser's built-in silent updater or download it directly from google.com/chrome. In our tests, autoplaying content that is muted still plays automatically. Autoplaying content with sound, whether it has visible controls or not, and whether it is set to play on loop or not, simply does not start playing. Note that this is all encompassing -- even autoplaying content you are expecting or is the main focus of the page does not play. YouTube videos, for example, no longer start playing automatically. And in case that's not enough, or if a page somehow circumvents the autoplaying block, you can still mute whole websites.

An anonymous reader quotes a report from Ars Technica: Google researchers have developed a deep-learning system designed to help computers better identify and isolate individual voices within a noisy environment. As noted in a post on the company's Google Research Blog this week, a team within the tech giant attempted to replicate the cocktail party effect, or the human brain's ability to focus on one source of audio while filtering out others -- just as you would while talking to a friend at a party. Google's method uses an audio-visual model, so it is primarily focused on isolating voices in videos. The company posted a number of YouTube videos showing the tech in action.

The company says this tech works on videos with a single audio track and can isolate voices in a video algorithmically, depending on who's talking, or by having a user manually select the face of the person whose voice they want to hear. Google says the visual component here is key, as the tech watches for when a person's mouth is moving to better identify which voices to focus on at a given point and to create more accurate individual speech tracks for the length of a video. According to the blog post, the researchers developed this model by gathering 100,000 videos of "lectures and talks" on YouTube, extracting nearly 2,000 hours worth of segments from those videos featuring unobstructed speech, then mixing that audio to create a "synthetic cocktail party" with artificial background noise added. Google then trained the tech to split that mixed audio by reading the "face thumbnails" of people speaking in each video frame and a spectrogram of that video's soundtrack. The system is able to sort out which audio source belongs to which face at a given time and create separate speech tracks for each speaker. Whew.

dmoberhaus writes: Last December, NASA announced that two new exoplanets had been hiding in plain sight among data from the Kepler space telescope. These two new planets weren't discovered by a human, however. Instead, an exoplanet hunting neural network -- a type of machine learning algorithm loosely modeled after the human brain -- had discovered the planets by finding subtle patterns in the Kepler data that would've been nearly impossible for a human to see. Last Thursday, Christopher Shallue, the lead Google engineer behind the exoplanet AI, announced in a blog post that the company was making the algorithm open source. In other words, anyone can download the code and help hunt for exoplanets in Kepler data.
Google's research blog called the December discovery "a successful proof-of-concept for using machine learning to discover exoplanets, and more generally another example of using machine learning to make meaningful gains in a variety of scientific disciplines (e.g. healthcare, quantum chemistry, and fusion research)."

Google has developed a "simple and consistent addressing system that works across India and globally." Called "Plus Codes," the location-based digital addressing system is designed for people with addresses that are not easily located through conventional descriptors like street names or house numbers. That's half of the world's urban population, according to a World Bank estimate. 9to5Google reports: Notably, this open source solution composed of 10 characters works globally and can be incorporated by other products and platforms for free, with a developer page available here. It works offline and on print when overlaid as a grid on existing maps. Places that are close together share similar plus codes, while the system is identifiable by the "+" symbol in every address. "This system is based on dividing the geographical surface of the Earth into tiny 'tiled areas,' attributing a unique code to each of them," reports Google. "This code simply comprises a '6-character + City' format that can be generated, shared and searched by anyone -- all that's needed is Google Maps on a smartphone."

The first four characters are the area code, describing a region of roughly 100 x 100 kilometers. The last six characters are the local code, describing the neighborhood and the building, an area of roughly 14 x 14 meters -- about the size of one half of a basketball court. The area code is not needed when navigating within a town, while another optional character can be appended to provide additional accuracy down to a 3 x 3 meter region. Users of Google Maps in India will be able to easily find the plus code for any area in the app, while the mapping service along with Search will support the entry of the new coordinate system. Plus codes for any location can also be found with this tool.

An anonymous reader quotes a report from VentureBeat: Google today launched Chrome 65 for Windows, Mac, Linux, and Android. Additions in this release include Material Design changes and new developer features. You can update to the latest version now using the browser's built-in silent updater or download it directly from google.com/chrome. Chrome 65 comes with a few visual changes. The most obvious is related to Google's Material Design mantra. The extensions page has been completely revamped to follow it. Next up, Chrome 65 replaces the Email Page Location link in Chrome for Mac's File menu with a Share submenu. As you might expect, Mac users can use this submenu to share the URL of a current tab via installed macOS Share Extensions. Speaking of Macs, Chrome 65 is also the last release for OS X 10.9 users. Chrome 66 will require OS X 10.10 or later. Moving on to developer features, Chrome 65 includes the CSS Paint API, which allows developers to programmatically generate an image, and the Server Timing API, which allows web servers to provide performance timing information via HTTP headers.

An anonymous reader quotes a report from Tom's Hardware: Google announced a 72-qubit universal quantum computer that promises the same low error rates the company saw in its first 9-qubit quantum computer. Google believes that this quantum computer, called Bristlecone, will be able to bring us to an age of quantum supremacy. In a recent announcement, Google said: "If a quantum processor can be operated with low enough error, it would be able to outperform a classical supercomputer on a well-defined computer science problem, an achievement known as quantum supremacy. These random circuits must be large in both number of qubits as well as computational length (depth). Although no one has achieved this goal yet, we calculate quantum supremacy can be comfortably demonstrated with 49 qubits, a circuit depth exceeding 40, and a two-qubit error below 0.5%. We believe the experimental demonstration of a quantum processor outperforming a supercomputer would be a watershed moment for our field, and remains one of our key objectives."

According to Google, a minimum error rate for quantum computers needs to be in the range of less than 1%, coupled with close to 100 qubits. Google seems to have achieved this so far with 72-qubit Bristlecone and its 1% error rate for readout, 0.1% for single-qubit gates, and 0.6% for two-qubit gates. Quantum computers will begin to become highly useful in solving real-world problems when we can achieve error rates of 0.1-1% coupled with hundreds of thousand to millions of qubits. According to Google, an ideal quantum computer would have at least hundreds of millions of qubits and an error rate lower than 0.01%. That may take several decades to achieve, even if we assume a "Moore's Law" of some kind for quantum computers (which so far seems to exist, seeing the progress of both Google and IBM in the past few years, as well as D-Wave).

Frederic Lardinois reports via TechCrunch: Hangouts Chat, Google's take on modern workplace communication, is now generally available and is becoming a core part of G Suite. Hangouts Chat was first announced at Google Cloud Next 2017, together with Hangouts Meet. While Meet went right into public availability, though, Chat went into an invite-only preview. Now, Google is rolling Chat out to all G Suite users over the course of the next seven days (so if you don't see it yet, don't despair). For all intents and purposes, Hangouts Chat is Google's take on Slack, Microsoft Teams and similar projects. Since Google first announced this project, Atlassian also joined the fray with the launch of Stride. Like its competitors, Chat is available on iOS, Android and the web.

Chat currently supports 28 languages and each room can have up to 8,000 members. What's maybe just as important, though, is that Google has already built an ecosystem of partners that are integrating with Chat by offering their own bots. They include the likes of Xero, RingCentral, UberConference, Salesforce, Zenefits, Zoom.ai, Jira, Trello, Wrike and Kayak. There's even a Giphy bot. Developers can also build their own bots and integrate their own services with Chat.

Google is rolling out tappable, visual stories that incorporate text, images, and videos in the style made popular by Snapchat. "It started widely testing the multimedia format, called AMP stories, today (Feb. 13) in an effort to help publishers engage more with readers on mobile," reports Quartz. Google announced the feature in a developer blog post. From the report: Users can now find Google stories in search results -- in a box called "visual stories" -- when they search on mobile at g.co/ampstories for the names of publishers that have begun using the format, such as CNN, Conde Nast, Hearst, Mashable, Meredith, Mic, Vox Media, and the Washington Post brands. Google worked with those publishers to develop the format. Desktop users can also get a taste of stories through Google's Accelerate Mobile Pages site. When a user selects a story, like Cosmopolitan magazine's piece on apple cider vinegar, it displays in a full-screen, slideshow format, similar to those on Snapchat and Instagram.

The multimedia format is part of Google's Accelerated Mobile Pages (AMP) project, a competitor to Facebook's Instant Articles that helps load pages faster on mobile devices. Like AMP, the AMP story format is open-sourced, so anyone can use it. However, Google is reportedly only displaying stories from a select group of publishers, including those it partnered with on the development, on its own site at the moment. The company said it plans to bring AMP stories to more Google products in the future, and expand the ways they appear in Google search.

Earlier today, YouTube barred Logan Paul from serving ads on his video channel in response to a "recent pattern of behavior" from him. Now, YouTube has announced a more formal and wider set of sanctions it's prepared to level on any creator that starts to post videos that are harmful to viewers, others in the YouTube community, or advertisers. TechCrunch reports: "We may remove a channel's eligibility to be recommended on YouTube, such as appearing on our home page, trending tab or watch next," Ariel Bardin, Vice President of Product Management at YouTube, writes in a blog post.

The full list of steps, as outlined by YouTube:
1. Premium Monetization Programs, Promotion and Content Development Partnerships. We may remove a channel from Google Preferred and also suspend, cancel or remove a creator's YouTube Original.
2. Monetization and Creator Support Privileges. We may suspend a channel's ability to serve ads, ability to earn revenue and potentially remove a channel from the YouTube Partner Program, including creator support and access to our YouTube Spaces.
3. Video Recommendations. We may remove a channel's eligibility to be recommended on YouTube, such as appearing on our home page, trending tab or watch next.

The changes are significant not just because they could really hit creators where it hurts, but because they also point to a real shift for the platform. YouTube has long been known as a home for edgy videos filled with pranks and potentially offensive content, made in the name of comedy or freedom of expression. Now, the site is turning over a new leaf, using a large team of human curators and AI to track the content of what's being posted, and these videos have a much bigger chance of falling afoul of YouTube's rules and getting dinged.

An anonymous reader quotes a report from Ars Technica: YouTube's latest strategy to fight the spread of misinformation involves putting a disclaimer on videos from certain news sources. The online video website announced it will start labeling videos posted by state-funded broadcasters to alert viewers that the content is, in some part, funded by a government source. YouTube will begin labeling videos today, and the policy extends to outlets including the US's Public Broadcasting Service (PBS) and the Russian government broadcaster RT. According to a report by The Wall Street Journal, PBS videos will now have the label "publicly funded American broadcaster," while RT will have this disclaimer: "RT is funded in whole or in part by the Russian government." The new policy is YouTube's way of informing viewers about where the content they're watching is coming from, a piece of information often hidden or left unsought by the viewers themselves. "The principle here is to provide more information to our users, and let our users make the judgment themselves, as opposed to us being in the business of providing any sort of editorial judgment on any of these things ourselves," YouTube Chief Product Officer Neal Mohan told the WSJ.

In a post on Google's Online Security Blog, two engineers described a novel chip-level patch that has been deployed across the company's entire infrastructure, resulting in only minor declines in performance in most cases. "The company has also posted details of the new technique, called Retpoline, in the hopes that other companies will be able to follow the same technique," reports The Verge. "If the claims hold, it would mean Intel and others have avoided the catastrophic slowdowns that many had predicted." From the report: "There has been speculation that the deployment of KPTI causes significant performance slowdowns," the post reads, referring to the company's "Kernel Page Table Isolation" technique. "Performance can vary, as the impact of the KPTI mitigations depends on the rate of system calls made by an application. On most of our workloads, including our cloud infrastructure, we see negligible impact on performance." "Of course, Google recommends thorough testing in your environment before deployment," the post continues. "We cannot guarantee any particular performance or operational impact."

Notably, the new technique only applies to one of the three variants involved in the new attacks. However, it's the variant that is arguably the most difficult to address. The other two vulnerabilities -- "bounds check bypass" and "rogue data cache load" -- would be addressed at the program and operating system level, respectively, and are unlikely to result in the same system-wide slowdowns.

An anonymous reader quotes a report from TechCrunch: In a blog post published minutes ago, Google's Security team announced what they have done to protect Google Cloud customers against the chip vulnerability announced earlier today. They also indicated their Project Zero team discovered this vulnerability last year (although they weren't specific with the timing). The company stated that it informed the chip makers of the issue, which is caused by a process known as "speculative execution." This is an advanced technique that enables the chip to essentially guess what instructions might logically be coming next to speed up execution. Unfortunately, that capability is vulnerable to malicious actors who could access critical information stored in memory, including encryption keys and passwords. According to Google, this affects all chip makers, including those from AMD, ARM and Intel (although AMD has denied they are vulnerable). In a blog post, Intel denied the vulnerability was confined to their chips, as had been reported by some outlets. The Google Security team wrote that they began taking steps to protect Google services from the flaw as soon as they learned about it.

schwit1 quotes ScienceAlert: In May 2017, researchers at Google Brain announced the creation of AutoML, an artificial intelligence (AI) that's capable of generating its own AIs. More recently, they decided to present AutoML with its biggest challenge to date, and the AI that can build AI created a 'child' that outperformed all of its human-made counterparts... For this particular child AI, which the researchers called NASNet, the task was recognising objects -- people, cars, traffic lights, handbags, backpacks, etc. -- in a video in real-time. AutoML would evaluate NASNet's performance and use that information to improve its child AI, repeating the process thousands of times.

When tested on the ImageNet image classification and COCO object detection data sets NASNet was 82.7 percent accurate at predicting images on ImageNet's validation set. This is 1.2 percent better than any previously published results, and the system is also 4 percent more efficient, with a 43.1 percent mean Average Precision (mAP).

The Google Pixel 2 and Pixel 2 XL both feature a custom Intel "Visual Core" co-processor, which is meant to improve speed and battery life when shooting photos with Google's HDR+ technology. The chip has been hanging out in the phone not really doing much of anything -- until now. TechCrunch reports of a new developer preview of Android 8.1 due out today that puts the chip to use. "The component is expected to further improve the handsets' cameras, which were already scoring good marks, production issues aside." From the report: According to the company, Pixel Visual Core has eight image processing unit (IPU) cores and 512 arithmetic logic units. Using machine learning, the company says it's able to speed things up by 5x, with one tenth of the energy. Access to the chip, combined with the Android Camera API means third-party photo apps will be able to take advantage of the system's speedy HDR+. Sounds swell, right? Of course, this is still just an early preview, only available to people who sign up for Google's Beta program. That means, among other things, dealing with potential bugs of an early build. Google wouldn't give us any more specific information with regards to when the feature will be unlocked for the public, but it's expected to arrive along with the 8.1 public beta in December.

YouTube will be implementing five new rules to protect minors from disturbing content. They include removing ads from inappropriate videos and blocking predatory comments from videos that feature minors. CNET reports: The new guidelines are as follows:

-Tougher application of Community Guidelines and faster enforcement through technology
-Removing ads from inappropriate videos targeting families
-Blocking inappropriate comments on videos featuring minors
-Providing guidance for creators who make family-friendly content
-Engaging and learning from experts

These rules follow recent reports that expose fundamental flaws in YouTube algorithms and screening protocol, which fail to recognize or pull down videos that feature disturbing imagery but are aimed at children for monetization purposes. Videos featuring children doing innocuous activities like exercising are also riddled with predatory or sexual comments from viewers, something YouTube is attempting to curb with its new guidelines.

In a blog post, Google's Product Manager, Chris Kleban, announced that the company is cutting the price of using Nvidia's Tesla GPUs through its Compute Engine by up to 36 percent. The older K80 GPUs will now cost $0.45 per hour while the more powerful P100 machines will cost $1.46 per minute (all with per-second billing). TechCrunch reports: The company is also dropping the prices for preemptible local SSDs by almost 40 percent. "Preemptible local SSDs" refers to local SSDs attached to Google's preemptible VMs. You can't attach GPUs to preemptible instances, though, so this is a nice little bonus announcement -- but it isn't going to directly benefit GPU users. As for the new GPU pricing, it's clear that Google is aiming this feature at developers who want to run their own machine learning workloads on its cloud, though there also are a number of other applications -- including physical simulations and molecular modeling -- that greatly benefit from the hundreds of cores that are now available on these GPUs. The P100, which is officially still in beta on the Google Cloud Platform, features 3594 cores, for example. Developers can attach up to four P100 and eight K80 dies to each instance. Like regular VMs, GPU users will also receive sustained-use discounts, though most users probably don't keep their GPUs running for a full month.

"Google today announced a forthcoming update to its Accelerated Mobile Pages, or AMP, web format that aims to discourage website owners from misusing the service," reports The Verge. "The company says that, starting in February 2018, AMP pages must contain content nearly identical to that of the standard page they're replicating." From the report: Currently, because AMP pages load faster and more clutter-free versions of a website, they naturally contain both fewer ads and less links to other portions of a site. That's led some site owners to publish two versions of a webpage: a standard page and an AMP-specific one that acts a teaser of sorts that directs users to the original. That original page, or canonical page in Google parlance, is by nature a slower loading page containing more ads and with a potentially lower bounce rate, which is the percentage of viewers who only view one page before leaving. Now, Google is cracking down on that behavior. "AMP was introduced to dramatically improve the performance of the web and deliver a fast, consistent content consumption experience," writes Ashish Mehta, an AMP product manager. "In keeping with this goal, we'll be enforcing the requirement of close parity between AMP and canonical page, for pages that wish to be shown in Google Search as AMPs."

Google is digging into the dark corners of the web to better secure people's accounts. From a report: For one year, Google researchers investigated the different ways hackers steal personal information and take over Google accounts. Google published its research, conducted between March 2016 and March 2017, on Thursday. Focusing exclusively on Google accounts and in partnership with the University of California, Berkeley, researchers created an automated system to scan public websites and criminal forums for stolen credentials. The group also investigated over 25,000 criminal hacking tools, which it received from undisclosed sources. Google said it is the first study taking a long term and comprehensive look at how criminals steal your data, and what tools are most popular. [...] Google researchers identified 788,000 potential victims of keylogging and 12.4 million potential victims of phishing. These types of attacks happen all the time. For example on average, the phishing tools Google studied collect 234,887 potentially valid login credentials, and the keylogging tools collected 14,879 credentials, each week.

YouTube has rolled out a new algorithm that the company says will more accurately reflect YouTube's guidelines for ad-friendly material and result in fewer videos being flagged as advertiser-friendly. "It will supposedly reduce the number of demonetized listings by 30 percent, so 'millions more videos' will be able to make money off the full range of advertisements," reports The Verge. From the report: A YouTube manager writes that the new algorithm was trained by nearly three months' worth of human reviews, starting after YouTube added a manual appeals process for creators in August. Theoretically, this should narrow the range of false positives -- videos that were incorrectly flagged for promoting drug use, using excessive profanity, highlighting gratuitous violence, or otherwise featuring content that advertisers might find objectionable. It's being applied retroactively, so creators who didn't appeal could still get some old videos remonetized. Google also encourages people to keep appealing potentially incorrect flags, because "this updated system is an improvement, but it's not perfect."

An anonymous reader writes: In a bid to help Android smartphone vendors the Linux LTS (Long Term Support) kernels will now be maintained for a period of six years. The Linux LTS initiative backed by the Linux Foundation has supported annual LTS kernels for two years worth of updates, but that is being changed for Linux 4.4+ at the request of Google and their Project Treble. This means the Linux 4.4 LTS kernel will be maintained through 2022 and the upcoming Linux 4.14 LTS through 2023 for security/bug fixes in order to last a complete "device lifecycle."

BrianFagioli shares a report from BetaNews: The documentation aspect of any project is very important, as it can help people to both understand it and track changes. Unfortunately, many developers aren't very interested in documentation aspect, so it often gets neglected. Luckily, if you want to maintain proper documentation and stay organized, today, Google is releasing its internal developer documentation style guide. This can quite literally guide your documentation, giving you a great starting point and keeping things consistent.

Jed Hartman, Technical Writer, Google says, "For some years now, our technical writers at Google have used an internal-only editorial style guide for most of our developer documentation. In order to better support external contributors to our open source projects, such as Kubernetes, AMP, or Dart, and to allow for more consistency across developer documentation, we're now making that style guide public. If you contribute documentation to projects like those, you now have direct access to useful guidance about voice, tone, word choice, and other style considerations. It can be useful for general issues, like reminders to use second person, present tense, active voice, and the serial comma; it can also be great for checking very specific issues, like whether to write 'app' or 'application' when you want to be consistent with the Google Developers style." You can access Google's style guide here.

Google announced in a blog post today that the Google Drive app for desktop will be shut down. The Verge reports: Support will be cut off on December 11th and the app will shut down completely on March 12th, 2018. Users who are still running the Drive app will start seeing notifications in October that it's "going away," and the company will steer customers towards one of two replacements depending on whether they're a consumer or business user. Google Drive the service isn't going anywhere. You can still access it from the web, smartphone apps, and either of the software options mentioned below. Google now has two fairly new software tools for backing up your data and/or accessing files in the cloud. There's Backup and Sync, the all-encompassing consumer app that replaces both the standalone Google Drive and Google Photos Uploader apps. It offers essentially the same functionality as Drive and works much the same way. And on the enterprise side, Google has rolled out Drive File Streamer, which saves space on your local drive while providing access to "all of your Google Drive files on demand, directly from your computer."

The latest version of Google Chrome has launched, bringing a host of new developer features like JavaScript modules and WebUSB support. An anonymous Slashdot reader shares a report from VentureBeat: Google has launched Chrome 61 for Windows, Mac, and Linux. Additions in this release include JavaScript modules and WebUSB support, among other developer features. You can update to the latest version now using the browser's built-in silent updater or download it directly from google.com/chrome. Google also released Chrome 61 for Android today. In addition to performance and stability fixes, you can expect two new features: Translate pages with a more compact toolbar and pick images with an improved image picker.

Chrome now supports JavaScript modules natively via the new element, letting developers declare a script's dependencies. Modules are already popular in third-party build tools, which use them to bundle only the required scripts. Native support means the browser can fetch granular dependencies in parallel, taking advantage of caching, avoiding duplications across the page, and ensuring the script executes in the correct order, all without a build step. Google recommends these two blog posts for more information: ECMAScript modules in browsers and ES6 Modules in Depth. Speaking of JavaScript, Chrome 61 also upgrades the browser's V8 JavaScript engine to version 6.1. Developers can expect performance improvements and a binary size reduction. The WebUSB API meanwhile allows web apps to access user-permitted USB devices. This enables all the functionality provided by hardware peripherals such as keyboards, mice, printers, and gamepads, while still preserving the security guarantees of the web.

"Researchers at Google have found a vulnerability in the way watermarks are used by stock imagery sites like Adobe Stock that makes it possible to remove the opaque stamp used to protect copyright," writes Khari Johnson via VentureBeat. "The consistent nature in which the watermarks are placed on photos can be exploited using an algorithm trained to recognize and automatically remove watermarks." From the report: Changing the position or opacity of a watermark do not impact the algorithm's ability to remove watermarks from images with copyright protection. Randomization, the researchers say, is required to keep images from being stolen. In results presented at the Computer Vision and Pattern Recognition conference last month, subtle modifications to each watermark can make it harder to remove watermarks. With these warped watermarks, attempts to get rid of watermarks with an algorithm or photo editing software leaves noticeable marks, rendering an image useless. "As often done with vulnerabilities discovered in operating systems, applications or protocols, we want to disclose this vulnerability and propose solutions in order to help the photography and stock image communities adapt and better protect its copyrighted content and creations," research scientists Tali Dekel and Michael Rubenstein wrote in a blog post today. "From our experiments much of the world's stock imagery is currently susceptible to this circumvention." You can learn more about the different types of randomization that can be done to combat watermark removal and see more example images in Google's blog post. The full report and research is available via the project's GitHub page.