"The Register has a story about a Windows backdoor that waits silently in memory for a 'magic packet' before springing into action," writes Slashdot reader fred133. "No outgoing traffic, just waiting..." From the report: Like a sleeper cell awaiting activation, a never-before-seen Windows backdoor dubbed Sleepwalker waits silently in memory for one specifically crafted network packet to wake it up and deliver commands using the malware's 23-instruction language. The commands can do everything from running code directly in memory to moving data off the computer. Malware researcher Dominik Reichel discovered the passive backdoor, which also has its own command language, and detailed Sleepwalker in a technical analysis on Monday. "What makes it worth writing up is what that packet carries: not a readable command, but a short program written in a command language of the backdoor's own design," Reichel said. "Its 23 instructions cover scheduling, several ways to move data, staged file delivery and running code directly in memory. Recovering the encryption key is not enough to understand one of these programs. The internal command language must be reverse engineered as well." In addition to having its own command language, it's also notable that the remote host can be a VMware VMCI target instead of a normal network address. "Taken as a whole, the approach here is consistent with a targeted, well-resourced operation rather than an opportunistic one," Reichel wrote. The malware, hidden inside a 64-bit Windows DLL file, impersonates Microsoft's dpapi.dll, part of Windows' data protection API for protecting sensitive data. It exports the same seven functions as the real dpapi.dll, but attempts to forward calls to a file named dpapisvc.dll, which is not a real Windows component. The file also has a forged ESET Management Agent version resource, and loads via side-loading into ERAAgent.exe, the Windows executable for ESET Management Agent. After confirming that its host process is named ERAAgent.exe, Sleepwalker goes to sleep inside the computer's memory, which also helps it remain hidden from traditional anti-virus tools. Unlike most backdoors, which call back to an attacker-controlled command-and-control (C2) server and start receiving commands, Sleepwalker lies in wait, checking every packet that passes through the network looking for a specific pattern - this is called a magic packet. Once it sniffs out a packet that matches the exact pattern, the backdoor decrypts the data and treats it as a command. "Because the backdoor never sends anything out on its own and does not open any obvious listening port by default, tools that watch for connections to known-bad domains or unusual outbound traffic will not see anything unusual," Reichel wrote. "The absence of outbound connections to known-bad infrastructure does not rule out an infection, either. A machine can be fully compromised by this backdoor while producing nothing at all for a network monitor to flag."

Read more of this story at Slashdot.

A developer says AliExpress is using the browser's WebAudio API to help fingerprint users by playing inaudible audio and measuring tiny differences in how their devices process it. CyberNews reports: The developer, "laserphile," wrote on their blog that they recently ran into some weird issues with their Bluetooth headphones. They couldn't play music via their phone when, at the same time, the AliExpress website was open on their PC. The headphones, laserphile explained, support multipoint Bluetooth audio so they can be connected to the PC and phone at the same time, for instance, playing music on the phone and announcing notifications through the PC. "Shortly after loading the AliExpress homepage, audio from my phone would stop playing. Closing the AliExpress tab fixes it immediately," the developer said in the blog post. "Muting the tab/Firefox/Windows does not help, and there is no visible video, music, or other media playing on the page. This seemed suspicious enough to investigate." It turns out that Alibaba has been secretly leveraging AliExpress users' audio systems to track them and build detailed fingerprints of them. [...] The AliExpress site was using the browser's WebAudio API to run invisible sound waves at zero volume. By measuring tiny hardware differences in how each PC processed those signals, the site created a unique digital fingerprint to track devices -- without user knowledge or consent. The secret audio path froze the developer's Bluetooth connection while covertly scraping hardware memory, screen dimensions, and network data in the background. The data collection extends beyond audio. Further inspection revealed that the same scripts also measure canvas, WebGL, hardware specs, WebRTC, mouse/touch events, and automation indicators. All of these form a broad device fingerprint that is sent back to Alibaba's telemetry servers. The simplest fix is to use a privacy-focused browser such as Firefox or Brave, which can limit or block this kind of fingerprinting. Brave goes further by randomizing fingerprint data and blocking the AliExpress tracking scripts involved.

Read more of this story at Slashdot.

An anonymous reader quotes a report from the BBC: The world's oceans are hotter than ever recorded, new data suggests, as they suffer from human-caused climate change and the growing El Nino weather phenomenon. The average surface temperature of the planet's seas outside the polar regions hit 21.1C (70F) on Saturday, according to figures from the European Copernicus climate change service. That edges past the 21.09C recorded on three separate days in March 2024, and is far above average for the time of year. [...] The data is based on sea temperatures 10m (32ft 10in) below the surface, using measurements from buoys, ships and satellites, which are combined to produce a global estimate. While the margin of record is currently very small and any global estimate comes with uncertainties, scientists say its timing is particularly notable. Average worldwide sea temperatures tend to reach their yearly peak in March or April, which corresponds to the end of summer in the southern hemisphere -- and not in August. The southern hemisphere contains more of the planet's ocean surface than the northern hemisphere and so exerts a bigger influence on average sea temperatures. What is especially concerning to scientists is that the oceans are already so hot when the natural El Nino weather phenomenon is still some way off its expected peak. "The fact that we are already breaking records is an early indicator of how strong the El Nino is becoming," said Dr Jeremy Grist, senior research fellow at the National Oceanography Centre in Southampton. "All things being equal we might expect the ocean temperature record to be broken again in March [or] April 2027," he added.

Read more of this story at Slashdot.

Amazon has raised the prices of its hardware devices by as much as 60%, blaming "significant increases in memory and storage component costs." TechCrunch reports: The price explosion impacted Fire TVs, Echos, Kindles, and Eeros. One egregious example that's been cited is that of the Echo Dot, one of Amazon's cheapest smart speakers, the price of which jumped 60% overnight, from $49.99 to $79.99. Price-tracking sites like CamelCamelCamel show the stark uptick, which has previously hovered much lower. [...] The company also said that it would continue to offer occasional promotions to customers over the course of the next year. Amazon said in a statement: "The consumer electronics industry is facing significant increases in memory and storage component costs. After absorbing these increases for as long as we could, we recently adjusted pricing across our product lines."

Read more of this story at Slashdot.

The SEC is investigating the near-collapse of AI-focused hedge fund Situational Awareness, sending subpoenas to major Wall Street banks for details about the fund's trades, borrowing, and communications with lenders. The fund, founded by former OpenAI researcher Leopold Aschenbrenner, managed over $30 billion and borrowed tens of billions more before leveraged bets unraveled, forcing it to sell most of its stock portfolio to Citadel at a discount. The New York Times reports: The subpoenas asked for details on the timing of Situational Awareness's trades and for its communications with lenders about the money it was borrowing, also known as "leverage," two of those people said. The subpoenas additionally warned the banks to preserve any information regarding the San Francisco hedge fund. The S.E.C. oversees financial markets with an eye toward protecting small investors, and has brought civil cases regularly against investment firms that produced large losses. Any investigation into Situational Awareness would be at its earliest stages, and it's no guarantee that it would lead to fines or other punishment. The hedge fund has not been accused of wrongdoing. [...] Situational Awareness had a fast rise and an even quicker retreat. Founded just two years ago by Leopold Aschenbrenner, a former researcher at OpenAI, it rode the A.I. boom to soaring investment returns. To achieve those results, however, the fund relied on heavy borrowing, as well as complicated and expensive financial instruments that magnify gains -- and losses. The latter piled up quickly last month when the stock prices of publicly traded, high-flying A.I. companies dipped. At the same time, shares in more traditional technology companies -- which the hedge fund had been betting against -- rose, compounding the problem. Situational Awareness was forced into a fire sale. It wound up selling most of its stock portfolio to a rival, Citadel, at a discount.

Read more of this story at Slashdot.

The Trump administration is proposing to make permanent a $103,265 fee on certain new H-1B visas, dramatically increasing costs for employers that rely on highly skilled foreign workers in tech, education, and research. "A federal judge in June ruled that the fee was illegal and blocked the Trump administration from collecting it," reports Reuters. "A Boston-based appeals court is reviewing that decision while a different court considers whether a judge properly rejected a challenge to the fee by a major business group." From the report: Trump's temporary fee increase expires in September, one year after it was issued. The proposed rule by the U.S. Department of Homeland Security, posted in the Federal Register on Monday, would make a fee of $103,265 permanent. It could be finalized by the end of the year. The H-1B program allows U.S. employers to hire foreign workers with training in specialty fields and offers 65,000 visas annually, with another 20,000 for workers with advanced degrees, approved for three to six years. Those visas typically came with fees between $2,000 and $5,000 before Trump's order. The fee would not apply to visas granted to foreign citizens already in the United States on student visas, who make up a large share of new H-1B recipients, or to renewals of current visas.

Read more of this story at Slashdot.

Raspberry Pi has published an official guide for building a cyberdeck, leaning into a viral trend of homemade portable computers built from repurposed cases, purses, and other oddball hardware. The guide uses a Raspberry Pi 5, keyboard, battery, and Pelican-style case. The Verge reports: Raspberry Pi's head of social, Ashley Whittaker, acknowledged the cyberdeck trend today, saying "we haven't been able to get away from cyberdecks this year." Raspberry Pi has also increased its prices multiple times this year due to rising costs for RAM and other components, which could be another reason to support this growing market of active and would-be hobbyists. The 16GB version of the Raspberry Pi 5 included in the company's official cyberdeck guide got a $100 price hike in April, nearly as much as the board's original $120 price.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: Taiwan has indicted nine people -- reportedly including an Nvidia senior manager and two Supermicro employees based in the country -- who allegedly helped forge documents to cover up illegal exports of high-end AI servers to China in violation of US export controls. On Monday, Reuters reported that prosecutors in Keelung did not release any names but confirmed that the suspects were "charged with breach of trust and document forgery." One suspect linked to a related case remains at large, accused of the "siphoning of funds from a distributor company involved" in helping China get access to restricted Nvidia chips. Since 2022, the US has required a license to export semiconductors to China, with the intent of protecting national security by staying ahead of China in the AI race. But after the US arrested Supermicro co-founder Wally Liaw in March for allegedly funneling $2.5 billion in restricted AI servers to China since 2024, Taiwan launched a separate probe to learn more about the scheme. Taiwan's investigation found that co-conspirators attempted to falsify documents to make it appear that 130 B300 servers were installed and in use at a Taiwan facility. But 74 of those servers "were ultimately exported and delivered to Chinese customers," Reuters reported, while the rest, 56, were intended to go to China but were blocked once Taiwanese customs officials "detected irregularities." According to Taiwan officials, Nvidia's manager and Supermicro employees "colluded with one another at various levels for enormous profit, illegally exporting high-end servers, increasing corporate compliance costs, and severely damaging our nation's international image." Supermicro said in a statement to PCMag: "Supermicro is committed to protecting our advanced technologies and intellectual property. We are working closely with Taiwanese authorities on recent events, which underscore the importance of continued collaboration across industry and government to strengthen safeguards, enhance supply chain visibility, and facilitate the enforcement of export control laws."

Read more of this story at Slashdot.

China has delayed its ambitious Chang'e 7 lunar mission until at least 2027 after scrubbing an August 24 launch. Poor weather conditions were likely a factor, with China's Manned Space Agency (CMSA) saying the decision reflects the need for "absolute safety." Scientific American reports: China and the U.S. are in a race to send humans back to the moon, a feat that that U.S. accomplished in the 1960s and that has never been achieved by any other country. Chang'e 7 is to play to a crucial part in China's ambitions: the probe is designed to search for frozen water in the Shackleton Crater, a pit that spans some 21 kilometers (13 miles) and sits beside the lunar south pole. No spacecraft has ever landed in that region, which scientists consider to be among the likeliest places to find water on the moon. Locating water that can be used by humans is vital to establishing any long-term lunar presence. NASA and China are eyeing the region to establish a crewed base -- and because this real estate is so precious, it matters who gets their first. Ars Technica notes in its reporting that the delayed launch was "most likely" due to Typhoon Narra, "which has 45 mph winds (equivalent to a moderate Tropical Storm) and continues to meander around in the waterway separating Hainan Island, where the spaceport is located, and mainland China." As for why the next launch window won't open until next February or March? "It likely is a combination of factors, including sunlight conditions in polar landing zones as well as the mission plan that requires a 90-day lunar orbit reconnaissance before landing," reports Ars. There's also a possibility, albeit an unlikely one, that "some technical issue was discovered at the last minute with either the complex Chang'e 7 payload, or the rocket."

Read more of this story at Slashdot.

The FDA has cleared the first Alzheimer's blood test for symptomatic patients as young as 40. The test measures amyloid and tau biomarkers from a single blood draw and identified positive biomarkers with 97.6% accuracy in a validation study. New Atlas reports: PrecivityAD2 is a single blood test that quickly assesses the proteins amyloid and tau, using mass spectrometry to measure plasma AB42/40 and p-tau217/np-tau217. The results are then scored, qualifying as negative, likely positive, or positive for AD. "Just over a year ago, there were no FDA-cleared blood tests for Alzheimer's, and now we have three," Isobel Coleman, chief executive officer of the Alzheimer's Drug Discovery Foundation (ADDF), said in a statement. "This is a proof point that sustained, early investment in translational science plays a catalytic role in moving promising ideas from the lab to patients." In a recent clinical validation study, 1,142 symptomatic adults were tested using PrecivityAD2. The process accurately identified 97.6% of positive biomarkers. What's more, the results remained consistent even in patients with comorbidities like chronic kidney disease and diabetes, which can distort plasma amyloid samples. It's worth noting that the test isn't intended for general health screenings, but for patients who already show some cognitive decline. "As PrecivityAD2 and tools like it move further into clinical practice, they will enable earlier detection, sharpen clinical trials, and build the foundation for precision medicine and combination therapies," Coleman added.

Read more of this story at Slashdot.

Two humanoid robots competing at the World Humanoid Robot Games in Beijing on Saturday ran the 100-meter dash faster than Usain Bolt's 9.58-second world-record time. "[T]he Tiangong Ultra humanoid robot completed the 100-meter dash in 9.39 seconds during a preliminary heat," reports Engadget. "Developed by Beijing Humanoid Robot Innovation Center, the record-holding robot overtook Honor's Lightning robot, which completed the sprint in 9.47 seconds. From the report: Despite the impressive finish times, don't expect these humanoid robots to be used for any commercial purposes yet. After their sprints, they slammed directly into pads and some even required human assistance to be stretchered off the course. However, it's still a major improvement from last year's inaugural World Humanoid Robot Games, where the Tiangong Ultra won the 100-meter dash in 21.5 seconds. Beijing has also hosted a humanoid robot half-marathon, where Honor's Lightning won in 50 minutes and 26 seconds, which also beat the pace of another human world record. "The World Humanoid Robot Games in Beijing kicked off August 22 and will run until the 26, hosting thousands of robot participants competing in more than just track and field events, including soccer, table tennis and boxing," adds Engadget.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Gizmodo: Dr. Dre, the pioneering rapper and seven-time Grammy-winning producer behind hit songs like Tupac's "California Love" and Eminem's "The Real Slim Shady," is very much pro using artificial intelligence tools to make music. In an interview with the New York Times, Dr. Dre and Interscope Records co-founder Jimmy Iovine spoke about artificial intelligence's recent foray into music production, with the producer characterizing those who oppose AI song generation tools as "afraid of learning new things." "I don't see it as a threat. I think the only people that see it as a threat are the people who have trouble creating," Dr. Dre said. "I had a discussion with a few people a few days ago. They were against AI, and I'm like, 'OK, you sound like the person that would have been against the drum machine when it came out.' Or synthesizers, right?" [...] Iovine, the music executive and Dr. Dre's partner in Beats by Dre, for his part, says he doesn't "see the downside at all." "There will be some crappy music. There's crappy music now," Iovine told the NYT. "In the studio, when gifted people have AI, they're going to make better records." Dr. Dre said he uses AI in producing "as a tool to see what it would do with what I just did." The award-winning producer also said he is definitely not alone among his peers in his stance on AI either: many producers use AI but fear to admit it. Iovine calls them "closet AI producers," and says Timbaland, the four-time Grammy-winning singer-producer behind a long list of hits like "Promiscuous" with Nelly Furtado, is also one.

Read more of this story at Slashdot.

Gluons are the quantum 'Gorilla Glue' involved in binding quarks together, ScienceAlert reminds us. But for nearly 50 years, physicists have hunted for "glueballs", exotic particles predicted by this theory of the strong interaction — the only type of particle in nature composed entirely of force mediators: Now, a collaboration working at a collider in Beijing says it has the clearest evidence yet that a known particle called X(2370) is dominated by one of these elusive states... Physicists sifted through the wreckage of high-energy particle collisions to find strong evidence of this glueball, made predominantly of gluons — massless force carriers of the strong interaction between quarks... The new evidence for this exotic, unstable glueball's ephemeral existence comes from a recent preprint on arXiv, presented at the International Conference on High Energy Physics in Brazil... This collaboration uses the Beijing Electron Positron Collider II, a particle accelerator and collider that smashes electrons into their antimatter counterparts, called positrons, to probe perplexing particle physics. The resultant glueball is an "unprecedented form of matter," says Jin Shan, a particle physicist at Nanjing University and one of the research team's leaders. "It not only enables the theory describing strong interactions to pass its most rigorous test, but also vastly expands the boundaries of our understanding of the physical world," he told Fan Chen at South China Morning Post... Previously, in a study published in the journal Physical Review Letters in 2024, the researchers analyzed 10 billion meson decay events. They measured the X(2370) particle's mass and spin parity — a property that dictates its interactions and almost-instantaneous decay — for the first time, finding complete agreement between their experimental values and predictions. Now, they've identified its additional decay modes and determined its flavor-singlet nature, "the most important characteristic of a glueball...." The authors describe their work as the clearest experimental result to emerge from nearly 50 years of glueball searches, supporting the long-standing theoretical prediction that gluons can bind together into a new form of matter. The discovery ""highlights advances in technology and analytical techniques that have recently revealed another decades-in-the-making physics mystery," the article points out. "More experiments are needed to further confirm and constrain the prototypes of the highly esoteric glueball; in other words, more particle smashing."

Read more of this story at Slashdot.

The long legal battle over ownership of Linux "is closer than ever to ending," reports the Register, "after a panel of three judges ruled a claim against IBM and Red Hat isn't valid, and that time has expired for further action." In 2021, an heir to SCO settled with IBM for $14.25 million — a sum that reflects the fact SCO had for years failed to produce strong evidence to back its claims. Another of SCO's legal successors, Xinuos, filed a new claim that IBM should be on the hook because Big Blue knew it did not own the code it contributed to Linux but instead had a non-exclusive license to use it. Xinuos argued that when IBM contributed Project Monterey code to Linux [25 years ago], it breached that license. Xinuos eventually took that argument to the US District Court for the Southern District of New York — and failed to convince it that IBM and Red Hat had a case to answer. Xinuos appealed, and on August 10th the United States Court of Appeals for the Second Circuit decided [PDF] not to revisit the District Court's decision, agreeing that the original legalese governing Project Monterey means it's too late to re-litigate the matter. The Appeals Court also agreed that Xinuos tried to frame the case as a licensing issue but failed, instead arguing that the issue was really about ownership. That ain't all, folks, because Xinuos intends to file a petition to have the case re-heard by the full bench of the Court of Appeals. That hardly ever happens, unless the court finds significant errors or major legal issues that make a rehearing worthwhile. Law firm Kaplan says the Second Circuit has allowed reviews of less than 0.03 percent of the cases it has handled. So perhaps this matter is now close to a final resolution. Back in 2000 Slashdot interviewed one of the presidents of SCO.

Read more of this story at Slashdot.

A Pew Research senior data scientist analyzed 500,000 English-language webpages from the past five years with an AI detection tool to estimate how many were likely written or edited by AI: Of all the pages in this sample, 10% show significant signs of AI authorship... If we filter old webpages out of our samples and look at only the pages published after the release of ChatGPT, the trend is even more pronounced. In the July 2026 snapshot, signs of AI authorship can be found in over one-third of pages published after ChatGPT was released. This is in line with other studies that have shown that large shares of recently published pages on the internet were likely written or substantially edited by AI. AI-authored text is not evenly spread across the web. When ChatGPT was first released, the kinds of linguistic patterns that can signal AI authorship appeared at similar rates across the main top-level web domains (.com, .org, .edu and .gov). But in samples from 2026, around one-in-ten pages with a .com domain show signs of AI authorship — about double the share on .org domains (4.6%), and 10 times the rate on .edu or .gov domains (both around 1%). They even identified specific ways online text has changed, "comparing the internet of today to a snapshot from 2023." Em dashes appear twice as often Oxford commas increased 63%. Words used more frequently by AI (like "delve," "interplay" or "testament") have more than doubled in usage. "Negative parallelism" comparisons ("it's not just X, it's Y") have nearly tripled.

Read more of this story at Slashdot.

3,346 backers pledged $1,198,844 on Kickstarter to build the (modular) handheld Linux computer, the Mecha Comet. It's slightly bigger and heavier than your smartphone, with a 4-inch AMOLED display and loaded up with things like microphones, speakers, a SIM tray, two USB ports and a mini HDMI port. There's three separate kill switches for wifi, microphone and camera — plus a "boot mode" switch for the SD card, and even three pins for instant access to its debug console. It's now expected to ship in October or November (and is still available for preorder). Snap-in extensions include a keyboard (with programmable colors for its backlit keys), a gamepad, and a special IO Breakout that exposes the 40 pogo pins for modules, sensors, and MCUs. They're all open hardware, and founder Shoaib Merchant says "I'm looking forward to more community designs over the upcoming months." He shows the screen in action in a video on Kickstarter, demonstrating native apps like Music, Notes, Settings, and Files — as well as a Terminal app running on its color screen. A Linux correspondent at How-To Geek notes its ships "with a custom, hardware-optimized Linux distro out of the box," so "it's like having a dedicated Linux terminal in your pocket." And they argue the new device "could be the future of Linux smartphones." I should also mention that you don't necessarily have to use the battery; it can power on over USB PD without one. Overall, you're getting the same I/O you'd expect from a computer in a smartphone-like form factor, complete with modular input devices — and it's running Linux under the hood. For me, that's the most innovative gadget I've seen in the Linux hardware space in a long time... It's powerful enough to run most Linux terminal apps, which opens up a lot of possibilities. You're also getting an NPU built right into the SoC, which opens up the possibility for on-device AI workflows — something not possible in budget smartphones. Now, I personally plan to use it for SSHing into my main computer and home lab and operating them remotely. That's also why I'll be putting the LTE modem in the M.2 slot rather than an SSD — mobile data means I can reach my machines even when there's no Wi-Fi around. I also showed it to my partner, and she sees this as the ideal minimalist phone for digital detoxing. There's a built-in media player, a basic notes app, and you can even run Spotify from the terminal if you want.

Read more of this story at Slashdot.

It's "radically different from other quantum computers" reports Phys.org. The Helios system uses trapped ions — charged atoms suspended in free space using electromagnetic fields. But it operates with 98 qubits — making it the largest trapped-ion quantum computer built so far. Quantinuum, the company behind the device, is based in Cambridge, U.K., and Broomfield, Colorado. It demonstrated earlier machines operating on 32 qubits in 2023 and 56 qubits in 2025... It relies on two major advances — one in hardware and one in software. First, the four-way X junction lets the system handle several tasks at once rather than one at a time, which is much faster. That was impossible in earlier QCCD [quantum charge-coupled device] machines, which could only move data back and forth in a single line or loop. Second, judicious use of the freedom offered by the two dimensions of the X junction relies on new classical control software called Helios runtime, which plans the smartest, fastest route for moving and processing the data. Together, these represent a substantial advance in the engineering of quantum computers. An important consequence is that Helios can perform computations that cannot be performed even on the largest supercomputers using known methods within reasonable amounts of time and power consumption. This demonstrates the ability of this device and its successors to surpass the computational prowess of classical computers — although their algorithms and hardware are rapidly advancing, too. Having said that, Helios' computations have just been random benchmark tests, so the practical importance of this leap forward is still limited. To perform quantum computations of practical importance in science and commerce, even the most optimistic estimates suggest that quantum computers must be enormously bigger and better — of the order of a million qubits.

Read more of this story at Slashdot.

A class action filed this week accused Amazon's Twitch gaming service of using video streams to fuel its AI products without licensing or obtaining permission, reports Engadget: Earlier this month, Twitch Support posted on X that it added an option to opt out of "having your channel content used to train generative AI content models across Amazon." Shortly after, Twitch hosted an episode of its recurring Patch Notes stream where Mike Minton, the company's chief product officer, said "if it was opt-in, nobody would opt in," when asked by a viewer why the Twitch settings for AI training aren't opt-in but instead enabled by default, requiring users to opt out if they don't want to participate. Now, the latest class action lawsuit claims that Twitch and Amazon committed a breach of contract when the companies decided to use its streamers as "free training stock for separate commercial AI products."

Read more of this story at Slashdot.

This week OpenAI launched a version of ChatGPT exclusively designed for teenagers between the ages of 13 and 17, reports CTV, "saying this version came with stronger protections around content related to mentions of suicide, self-harm and romantic or sexual chats." [Tech analyst Carmi Levy says] "What it's doing is it's using cues, it's sort of watching how you're using the tool, and it's trying to guess your age." Levy said if ChatGPT thinks a person has lied about their age when they first signed up, it will automatically put them into the Teens mode for protection. "It includes those sort of two things: prevent you from having experiences that are age inappropriate, and also add features that are aligned with academics, so that makes it easier to study, makes it less likely for you to cheat on an assignment, provide resources that that that allow you to learn more effectively," he said. Levy reminded users that this is still an application and a web service, which means it's constantly collecting personal information about the users' activities on the device and using the data to estimate their age... In the event a user is talking about self-harm, this new version will block the conversation and suggest additional help or call emergency or mental health services, Levy explained. It would do the same with violent content, eating disorders, engaging in dangerous activities, as well as explicit, sexual or graphic content. "It's using AI again to sort of parse out that conversation: 'are we headed in a in a significant direction? If so, throw up that guardrail, stop it from happening, redirect.'" However, the tech analyst said a number of people who have attempted to bypass the guiderails for research have succeeded in doing so "with a little bit of prompting.... I wouldn't call it 100% safe. That's kind of the catch-22 of the entire AI industry." This will lead parents to think that their children are safer and reduce the oversight on their online activities, which Levy described as a "perverse impact" that is "problematic." "Parents should not assume that OpenAI has their kids' best interests at heart. They should also not assume that these technologies are perfect. They will miss certain things, and they will also engage in what we call false positives," he said. "An adult who's legitimately is in the platform might be tagged as a child because the AI simply got it wrong." OpenAI's blog post says the new mode was designed for "preserving the ability to learn, create, and explore." But CNN notes the move "comes as OpenAI faces a string of lawsuits alleging that inappropriate conversations with ChatGPT contributed to the harm or deaths of young people." [New tools] include nudges for teen users to take a break if they've been active for 90 minutes within a three-hour window. OpenAI is also adding options for teens or parents to set "Quiet Hours" when ChatGPT is unavailable or "Study Hours" where ChatGPT's study mode for learning will be on by default. The company also said it has improved protections to prevent ChatGPT from expressing personal feelings toward a user. And it's adding reminders that caution teens to think twice if they try to upload a private or sensitive image to the chatbot... Nearly a third of American teens use AI chatbots daily, a Pew Research Center study published in December found. ChatGPT was by far the most popular AI chatbot in the survey, with more than half of those teens reporting having used it. Engadget shared this observation from Josh Golin, the executive director of Fairplay, a non-profit advocating for policies to make the internet safer for children. "Most parents, overwhelmingly, do not avail themselves of [parental oversight] tools, in part because they are deliberately made hard to find and use." Golin's group wanted to see testing of the tools first to see whether they led to better outcomes for teenagers. "The way this was just announced feels more like a response aimed at public relations than actual deep thinking about what kind of support [teens] need." And Mashable got a similar quote from Robbie Torney, head of AI/digital assessments for Common Sense Media's Youth AI Safety Institute. OpenAI's teen experience makes important safety commitments, but "What we really need now is evidence that they work, and we need this evidence, including our testing and the testing of others, before anyone, including parents, puts real trust in ChatGPT for Teens." Importantly, Torney said parents shouldn't assume their child is using ChatGPT for Teens. While OpenAI uses age estimation to place users in the age-appropriate version of ChatGPT, it can still mistake a user as older. Parents can be sure their child is using the teen version by linking their ChatGPT account to their child's and implementing OpenAI's parental control features. If a teen has previously lied about their age — and plenty do — they may continue to access the adult version of ChatGPT... Torney said caregivers must enact parental controls to receive notifications from OpenAI should their teen discuss suicide, self-harm, or acts of violence in immediate or dangerous ways... An OpenAI spokesperson said every notification is reviewed by a trained human before it's sent. The company strives to review notifications in under one hour.

Read more of this story at Slashdot.

"Iran shut down a British power plant for four days in an unprecedented cyber attack," reports the Telegraph. More details from the BBC: The government said that at no point was there a risk to the UK's energy system, but the Department for Energy Security and Net Zero (DESNZ) has contacted power companies to advise them about the risk of cyber attacks... DESNZ said the incident had affected a small-scale generator and at no point had there been a risk to the wider energy system. The UK's power network has a number of smaller gas generators which provide short-term power when needed. Thanks to Alain Williams (Slashdot reader #2,972) for sharing the news.

Read more of this story at Slashdot.