Stability AI has announced it would allow artists to remove their work from the training dataset for an upcoming Stable Diffusion 3.0 release. From a report: The move comes as an artist advocacy group called Spawning tweeted that Stability AI would honor opt-out requests collected on its Have I Been Trained website. The details of how the plan will be implemented remain incomplete and unclear, however. As a brief recap, Stable Diffusion, an AI image synthesis model, gained its ability to generate images by "learning" from a large dataset of images scraped from the Internet without consulting any rights holders for permission. Some artists are upset about it because Stable Diffusion generates images that can potentially rival human artists in an unlimited quantity.

Google this week released OSV-Scanner -- an open source vulnerability scanner linked to the OSV.dev database that debuted last year. From a report: Written in the Go programming language, OSV-Scanner is designed to scan open source applications to assess the security of any incorporated dependencies -- software libraries that get added to projects to provide pre-built functions so developers don't have to recreate those functions on their own. Modern applications can have a lot of dependencies. For example, researchers from Mozilla and Concordia University in Canada recently created a single-page web application with the React framework using the create-react-app command. The result was a project with seven runtime dependencies and nine development dependencies.

But each of these direct dependencies had other dependencies, known as transitive dependencies. The react package includes loose-envify as a transitive dependency -- one that itself depends on other libraries. All told, this basic single-page "Hello world" app required a total of 1,764 dependencies. As Rex Pan, a software engineer on Google's Open Source Security Team, observed on Tuesday in a blog post, vetting thousands of dependences isn't something developers can do on their own.

An anonymous reader quotes a report from Reuters: Meta has halted construction of two data centers in Odense, Denmark, and will instead focus on a new type of data center used for artificial intelligence (AI), a spokesperson said on Thursday. Facebook-owner Meta already has two large data centers in Odense, but only one of the three other centers currently under development there will be completed. Construction on the two halted data centres in Odense began in August. However, on Tuesday Meta terminated the deal with contracting company Per Aarsleff worth 2.4 billion Danish crowns ($344 million). "Over the past month, we have announced a number of measures to make us a more streamlined organization," Meta spokesperson Peter Munster told Reuters. "A significant part of these measures is to shift a larger part of our resources to high-priority growth areas, including a strategic investment in artificial intelligence," he said.

The company's traditional data centers house servers for apps such as Facebook and Instagram. But the calculations needed for AI require a new generation of data centers, Munster said.

Slashdot reader hackingbear writes: Peter Wennink, the chief executive of ASML Holding NV, the Dutch semiconductor equipment maker, on Tuesday questioned whether a U.S. push to get the Netherlands to adopt new rules restricting exports to China make sense. "He said that following U.S. pressure, the Dutch government has already restricted ASML from exporting its most advanced lithography machines to China since 2019, something he said has benefited U.S. companies selling alternative technology," reports Reuters. "He said that while 15% of ASML's sales are in China, at U.S. chip equipment suppliers 'it is 25 or sometimes more than 30%.'"

In response to U.S. claims that advanced chips owned by China pose a threat to national security due to military applications and the rise of artificial intelligence, Wennink said: "What constitutes national security is for Americans to determine. But it is common knowledge that chip technology for purely military applications is usually ten, fifteen years old. The technology used to make such chips can still be sold to China. Artificial intelligence requires the most advanced chips. They are made with EUV and are therefore not produced in China. But those chips are simply sold, also to the Chinese. American chip manufacturers have no problem with China as a customer."

Apple is developing a 15.5-inch MacBook Air that could launch in the spring of 2023, according to display analyst Ross Young. MacRumors reports: In a tweet shared with super followers, Young said that production on panels designed for the MacBook Air will start in the first quarter of 2023. A 15.5-inch MacBook Air will be sized between the 14 and 16-inch MacBook Pro models, and it will be the largest MacBook Air to date. The current model measures in at 13.6 inches. While Young now says that the display will be 15.5 inches, he previously said it would be around 15.2 inches.

Apple's rumored 15-inch MacBook Air is expected to feature the same general design as the 13-inch MacBook Air that was released in 2022 with flat edges, a large Force Touch trackpad, a keyboard with function keys, and more. It will also likely include a MagSafe charging port, upgraded speaker system, and a 1080p camera. As for chips, the larger-screened MacBook Air could be available with M2 and M2 Pro chip options, and in comparison to the MacBook Pro models, it is not expected to feature the same mini-LED display or ProMotion technology.

TikTok recommends self-harm and eating disorder content to some users within minutes of joining the platform, according to a new report published Wednesday by the Center for Countering Digital Hate (CCDH). CBS News: The new study had researchers set up TikTok accounts posing as 13-year-old users interested in content about body image and mental health. It found that within as few as 2.6 minutes after joining the app, TikTok's algorithm recommended suicidal content. The report showed that eating disorder content was recommended within as few as 8 minutes.

Over the course of this study, researchers found 56 TikTok hashtags hosting eating disorder videos with over 13.2 billion views. "The new report by the Center for Countering Digital Hate underscores why it is way past time for TikTok to take steps to address the platform's dangerous algorithmic amplification," said James P. Steyer, Founder and CEO of Common Sense Media, which is unaffiliated with the study. "TikTok's algorithm is bombarding teens with harmful content that promote suicide, eating disorders, and body image issues that is fueling the teens' mental health crisis." The CCDH report details how TikTok's algorithms refine the videos shown to users as the app gathers more information about their preferences and interests. The algorithmic suggestions on the "For You" feed are designed, as the app puts it, to be "central to the TikTok experience." But new research shows that the video platform can push harmful content to vulnerable users as it seeks to keep them interested. Further reading: For teen girls, TikTok is the 'social media equivalent of razor blades in candy,' new report claims

The New York State Department of Financial Services (NYDFS) on Thursday issued digital asset guidance to state-regulated banks laying out what information financial institutions must submit before getting approval to engage in virtual currency-related activities. From a reportL: The guidance, one of the clearest paths forward yet for banks to offer cryptocurrency services, instructs banks to submit a business plan with details of the proposed activity, detail how such a service would impact the bank's capital and liquidity and inform NYDFS of its plans at least 90 days beforehand.

In a statement, NYDFS Superintendent Adrienne Harris said the new policies are "critical to ensuring that consumers' hard-earned money is protected" and that New York-regulated banks remain competitive. The regulator will "make a comprehensive assessment" of the information presented under the guidance to determine whether a bank should be permitted to engage in a proposed crypto-related activity, according to an industry letter sent Thursday to regulated institutions.

Jamming and spoofing attacks can cripple GPS connections entirely or make something appear in the wrong location, causing disruption and safety issues. Just ask Russia. From a report: New data analysis reveals that multiple major Russian cities appear to have faced widespread GPS disruption during the past week. The signal interference follows Ukraine launching long-range drone attacks deep into Russian territory, and it may act as a way to potentially stop drones that rely upon GPS for navigation, experts say. The GPS interference has "expanded on a scale that hasn't been seen before," says Erik Kannike, a program manager at Estonian defense intelligence firm SensusQ who has been monitoring the situation. "What we're seeing now, since about a week ago, is GPS jamming bubbles covering hundreds if not thousands of kilometers around tactical cities."

The GPS issues were first spotted by the monitoring system GPSJam, which uses data from planes to track problems with the satellite navigation system. The website has logged an increasing number of GPS disturbances in the Russian cities of Saratov, Volgograd, and Penza since the start of December. All of the cities are in Eastern Russia and within hundreds of kilometers of the border with Ukraine. On December 5, GPSJam logged a limited amount of GPS interference in Russia -- the majority of registered interference took place around Moscow, where the Kremlin for years has tampered with GPS connections. However, since December 11, multiple areas of the country have faced GPS disruption, data gathered by GPSJam shows. In addition, wireless data analytics firm Aurora Insight measured an increase in GPS signal levels in the area at the start of December -- a sign that potential GPS interference could have happened.

The Senate has passed legislation to ban TikTok from US government devices, in a move designed to limit perceived information-security risks stemming from the social media app. From a report: The vote by unanimous consent approved the No TikTok on Government Devices Act, a bill authored by Missouri Republican Sen. Josh Hawley. The move marks lawmakers' latest step against the short-form video app that has become popular with over a billion users worldwide. US officials fear that TikTok's user data could end up in the hands of the Chinese government due to that country's influence over TikTok's parent, ByteDance.

A companion bill was introduced in the House last year by Colorado Republican Rep. Ken Buck. It has yet to be approved by members of the House Oversight Committee. House Speaker Nancy Pelosi said Thursday it isn't yet clear whether the chamber will take up the TikTok bill in light of its Senate passage, saying lawmakers were consulting with White House officials on its language.

The Linux Foundation has partnered with some of the world's biggest technology companies to develop interoperable and open map data, in what is a clear move to counter Google's dominance in the mapping realm. From a report: The Overture Maps Foundation, as the new effort is called, is officially hosted by the Linux Foundation, but the program is driven by Amazon Web Services (AWS), Facebook's parent company Meta, Microsoft, and Dutch mapping company TomTom. The ultimate mission of the Overture Maps Foundation to power new map products through openly available datasets that can be used and reused across applications and businesses, with each member throwing their own data and resources into the mix.

"Mapping the physical environment and every community in the world, even as they grow and change, is a massively complex challenge that no one organization can manage," noted the Linux Foundation's executive director Jim Zemlin in a press release. "Industry needs to come together to do this for the benefit of all." Map and location data plays such a fundamental role across society today, powering everything from IoT (internet of things) devices and self-driving cars, to logistics and big data visualization tools. Having all that data under the auspices of just one or two mega-firms can be hugely restrictive in terms of what companies can do with the data and what features they have at their disposal, not to mention the costs involved in licensing it.

Bloomberg: Last March, brain-computing interface expert Tom Oxley sat down to dinner with Amazon founder Jeff Bezos to talk about Oxley's nascent company, Synchron. That meal in Ojai, Calif., ended with something better than dessert: Bezos told Oxley that he wanted to invest in the business. Synchron said Thursday it completed a $75 million funding round, part of it from Bezos Expeditions. The financing was led by ARCH Venture Partners, and includes a check from Gates Frontier, the venture investment arm of Microsoft co-founder Bill Gates, and others. Existing investors also participated, including Khosla Ventures -- whose founder, Vinod Khosla, introduced Oxley to Gates.

Brain-computer interfaces, known as BCIs, can interpret and stimulate parts of the brain and are seen as a possible treatment for brain injuries. New investors approached Synchron "through the lens of making an impact in neurology in an area of need," Oxley said in an interview. They "saw BCI as a future therapeutic." About 100 million people globally have upper limb impairment, he said, and could benefit from the technology. [...] Synchron's Switch device aims to help paralyzed people, such as those with amyotrophic lateral sclerosis (ALS), communicate by controlling computer cursors with their minds. The company has already enrolled three patients in a six-person US feasibility trial and implanted the device in two of them. Oxley expects the remaining patients to enroll and receive their implants in the next few months.

Microsoft has quietly banned cryptocurrency mining from its online services, and says it did so to protect all customers of its clouds. From a report: The Windows and Azure titan slipped the prohibition into an update of its Universal License Terms for Online Services that came into effect on December 1. That document covers any "Microsoft-hosted service to which Customer subscribes under a Microsoft volume licensing agreement," and on The Register's reading, mostly concerns itself with Azure.

Microsoft's Summary of Changes to the license states: "Updated Acceptable Use Policy to clarify that mining cryptocurrency is prohibited without prior Microsoft approval." Within the license itself there's hardly any more info. A section headed "Acceptable Use Policy" states: "Neither Customer, nor those that access an Online Service through Customer, may use an Online Service: to mine cryptocurrency without Microsoft's prior written approval."

An anonymous reader quotes a report from The Guardian: Rules around taking liquids and laptops through airport security will be eased from June 2024, the government has said. The announcement of the biggest relaxation of aviation security regulations in decades confirms reports last month that the change would come in the year after next. Passengers at most major UK airports will be able to carry liquids in containers holding up to two liters, a huge increase from the current limit of 100ml. Travelers will also no longer need to carry the containers in clear plastic bags, or remove tablets and laptops from hand luggage at checkpoints. The Department for Transport said major airports would be required to install new technology that gives security staff more detailed images of what is in passengers' bags. It will lay new legislation around the changes in parliament on Thursday. The transport secretary, Mark Harper, said: "The tiny toiletry has become a staple of airport security checkpoints, but that's all set to change. I'm streamlining cabin bag rules at airports while enhancing security."

"By 2024, major airports across the UK will have the latest security tech installed, reducing queueing times, improving the passenger experience, and most importantly detecting potential threats."

The US seized dozens of internet domains and charged six people in a sting intended to bring down a network of cyberattack-for-hire services, the Department of Justice announced on Wednesday. Bloomberg reports: In all, the US obtained a court order to seize 48 websites, and six people were criminally charged in relation to the takedowns, according to federal prosecutors. The FBI was in the process of seizing the websites, officials said Wednesday. The websites were used to launch, or attempt to launch, millions of so-called DDoS attacks around the world, the DOJ said in a statement. Short for distributed-denial-of- service, DDoS attacks direct huge amounts of junk internet traffic at a website or computer network to knock it offline.

DDoS-for-hire services often refer to themselves as "stresser" or "booter" tools that purport to offer a way for individuals to test the resilience of websites and services they operate, according to cybersecurity experts. In reality, the services are often used for harassment, extortion and criminal mischief, they say. The sites seized by the FBI include royalstresser, securityteam and dragonstresser, among others.

A new patent granted to Apple suggests the company could use satellite communications for more than just getting help in an emergency. 9to5Mac reports: Emergency SOS via Satellite was one of the headline features of September's Apple event -- so much so that the Far Out event name referenced it. The service launched in the US and Canada last month, and was yesterday extended to the UK, France, Germany, and Ireland. More countries will follow. A patent granted on the same day the service expanded to more countries suggests that Apple satellite plans may extend beyond text, and beyond emergency use.

Patently Apple spotted it: "Satellite communications data conveyed by transceivers #28 and antenna radiators #30 may include media data (e.g., streaming video, television data, satellite radio data, etc.), voice data (e.g., telephone voice data), internet data, and/or any other desired data." Apple has currently committed $450M to support the satellite communications feature, a reasonably sizeable amount of money even by Apple standards for a service that will be of use to a tiny fraction of iPhone owners. But if it's the start of something more, then the investment could look rather modest.

U.S. prosecutors on Wednesday said they have charged eight individuals in a securities fraud scheme, alleging they reaped about $114 million from by using Twitter and Discord to manipulate stocks. Reuters reports: The eight men allegedly purported to be successful traders on the social media platforms and then engaged in a so-called "pump and dump" scheme by hyping particular stocks to their followers with the intent to dump them once prices had risen, according to prosecutors in the Southern District of Texas.

The U.S. Securities and Exchange Commission (SEC) said it has filed related civil charges against the defendants in the scheme, claiming that seven of the defendants used Twitter and Discord to boost stocks. It said the eighth was charged with aiding and abetting the scheme with his podcast. The individuals charged were Texas residents Edward Constantinescu, Perry Matlock, John Rybarczyk and Dan Knight, along with California residents Gary Deel and Tom Cooperman, Stefan Hrvatin of Miami and Mitchell Hennessey of Hoboken, New Jersey.

Twitter co-founder Jack Dorsey said in a blog post on Tuesday that he will give a grant of $1 million per year to encrypted messaging app Signal, the first in a series of grants he plans to make to support "open internet development." Reuters reports: Social media should not be "owned by a single company or group of companies," and needs to be "resilient to corporate and government influence," Dorsey wrote in a post on Revue, a newsletter service owned by Twitter. [Editor's note: The post has been moved to Pastebin since Revue is shutting down early next year.] TechCrunch adds: Dorsey said that his hope to build a Twitter according to his wishes died in 2020 with the entrance of an unnamed activist investor. "I planned my exit at that moment knowing I was no longer right for the company," he wrote. The principles he had hoped to build on -- resilience to corporate and government control, user-controlled content with no exceptions and algorithmic moderation -- are not present in today's Twitter, nor in the one he led, he admitted. Even so, he wrote that, contrary to the insinuations accompanying the so-called Twitter Files, "there was no ill intent or hidden agendas, and everyone acted according to the best information we had at the time."

As to actual solutions, Dorsey is of course hard at work (or at least present) at Bluesky, but he calls out Mastodon and Matrix as other worthwhile avenues for development: "There will be many more. One will have a chance at becoming a standard like HTTP or SMTP. This isn't about a 'decentralized Twitter.' This is a focused and urgent push for a foundational core technology standard to make social media a native part of the internet."

A new feature in Firefox version 108 that may please musicians is the improved support for the Web MIDI API. "The MIDI standard is very close to a remarkable 40 years old, and Web MIDI does just what the name implies: it allows web apps to send and receive MIDI signals to and from musical instruments," reports The Register. "In principle this will allow sequencer apps to be implemented in Javascript." From the report: Amusingly, the last time The Reg mentioned Web MIDI, it was because Apple was taking it off Safari users, allegedly because of security concerns. Firefox 108 addresses that with a new security mechanism for preventing, and optionally permitting, apps inside browser tabs to access hardware resources -- in this instance, your MIDI ports. No, this does not mean that you can listen to CANYON.MID directly within Firefox. .MID files are not the same as General MIDI. But if you are nostalgic for that for some reason, help is at hand. A full list of features and changes can be found here.

An anonymous reader quotes a report from Ars Technica: YouTube has announced a plan to crack down on spam and abusive content in comments and livestream chats. Of course, YouTube will be doing this with bots, which will now have the power to issue timeouts to users and instantly remove comments that are deemed abusive. YouTube's post says, "We've been working on improving our automated detection systems and machine learning models to identify and remove spam. In fact, we've removed over 1.1 billion spammy comments in the first six months of 2022." It later adds, "We've improved our spambot detection to keep bots out of live chats."

When YouTube removes a message, the company says it will warn the poster that the message has been removed. The company adds, "If a user continues to leave multiple abusive comments, they may receive a timeout and be temporarily unable to comment for up to 24 hours." [...] It does not appear that YouTube is involving channel owners in any of these moderation decisions. Note that the post says YouTube will warn the poster (not the channel owner) of automated content removal and that if users disagree with the automated comment removal, they can "submit feedback" to YouTube. The "submit feedback" link on many Google products is a black hole suggestion box and not any kind of comment moderation queue, so it sounds like there will be no one that responds to a moderation dispute. YouTube says this automatic content moderation will only delete comments that violate the community guidelines—a list of pretty basic content bans—so hopefully it will stick to that.

Users of the popular MetaMask Web3 wallet will soon be able to buy the second-most-popular cryptocurrency, Ether, via PayPal. From a report: Similar to PayPal's checkout feature at online stores like Etsy and eBay, the integration with ConsenSys's MetaMask will let users buy and transfer Ether by logging in into MetaMask, tapping the "buy" button, and logging into PayPal before making a purchase.

The company said in a statement that select U.S.-based MetaMask customers will be able to use PayPal to buy Ether as of Wednesday, and that the feature will be rolled out to all U.S. users in the coming weeks. A crypto wallet like MetaMask is often the starting point for interacting with Web3 applications like play-to-earn games and some metaverse platforms. Adding PayPal to MetaMask could broaden the customer base for some of these applications by removing the complexity from buying crypto.