An anonymous reader shared this report from The Verge: In early December, I brought you the news that Google has begun replacing Verge headlines, and those of our competitors, with AI clickbait nonsense in its content feed [which appears on the leftmost homescreen page of many Android phones and the Google app's homepage]. Google appeared to be backing away from the experiment, but now tells The Verge that its AI headlines in Google Discover are a feature, one that "performs well for user satisfaction." I once again see lots of misleading claims every time I check my phone...

For example, Google's AI claimed last week that "US reverses foreign drone ban," citing and linking to this PCMag story for the news. That's not just false — PCMag took pains to explain that it's false in the story that Google links to...! What does the author of that PCMag story think? "It makes me feel icky," Jim Fisher tells me over the phone. "I'd encourage people to click on stories and read them, and not trust what Google is spoon-feeding them." He says Google should be using the headline that humans wrote, and if Google needs a summary, it can use the ones that publications already submit to help search engines parse our work.

Google claims it's not rewriting headlines. It characterizes these new offerings as "trending topics," even though each "trending topic" presents itself as one of our stories, links to our stories, and uses our images, all without competent fact-checking to ensure the AI is getting them right... The AI is also no longer restricted to roughly four words per headline, so I no longer see nonsense headlines like "Microsoft developers using AI" or "AI tag debate heats." (Instead, I occasionally see tripe like "Fares: Need AAA & AA Games" or "Dispatch sold millions; few avoided romance.")

But Google's AI has no clue what parts of these stories are new, relevant, significant, or true, and it can easily confuse one story for another. On December 26th, Google told me that "Steam Machine price & HDMI details emerge." They hadn't. On January 11th, Google proclaimed that "ASUS ROG Ally X arrives." (It arrived in 2024; the new Xbox Ally arrived months ago.) On January 20th, it wrote that "Glasses-free 3D tech wows," introducing readers to "New 3D tech called Immensity from Leia" — but linking to this TechRadar story about an entirely different company called Visual Semiconductor...

Google declined our request for an interview to more fully explain the idea.
The site Android Police spotted more inaccurate headlines in December: A story from 9to5Google, which was actually titled 'Don't buy a Qi2 25W wireless charger hoping for faster speeds — just get the 'slower' one instead' was retitled as 'Qi2 slows older Pixels.' Similarly, Ars Technica's 'Valve's Steam Machine looks like a console, but don't expect it to be priced like one' was changed to 'Steam Machine price revealed.' At the time, we believed that the inaccuracies were due to the feature being unstable and in early testing.... Now, Google has stopped calling Discover replacing human-written headlines as an "experiment."
"Google buries a 'Generated with AI, which can make mistakes' message under the 'See more' button in the summary," reports 9to5Google, "making it look like this is the publisher's intended headline." While it is obvious that Google has refined this feature over the past couple of months, it doesn't take long to still find plenty of misleading headlines throughout Discover... Another article from NotebookCheck about an Anker power bank with a retractable cable was given a headline that's about another product entirely. A pair of headlines from Tom's Hardware and PCMag, meanwhile, show the two sides of using AI for this purpose. The Tom's Hardware headline, "Free GPU & Amazon Scams," isn't representative of the actual article, which is about someone who bought a GPU from Amazon, canceled their order, and the retailer shipped it anyway. There's nothing about "Amazon Scams" in the article.

Richard Stallman spoke Friday at Atlanta's Georgia Institute of Technology, continuing his activism for free software while also addressing today's new technologies.

Speaking about AI, Stallman warned that "nowadays, people often use the term artificial intelligence for things that aren't intelligent at all..." He makes a point of calling large language models "generators" because "They generate text and they don't understand really what that text means." (And they also make mistakes "without batting a virtual eyelash. So you can't trust anything that they generate.") Stallman says "Every time you call them AI, you are endorsing the claim that they are intelligent and they're not. So let's let's refuse to do that."

"So I've come up with the term Pretend Intelligence. We could call it PI. And if we start saying this more often, we might help overcome this marketing hype campaign that wants people to trust those systems, and trust their lives and all their activities to the control of those systems and the big companies that develop and control them."

"By the way, as far as I can tell, none of them is free software."

When it comes to today's cars, Stallman says they contain "malicious functionalities... Cars should not be connected. They should not upload anything." (He adds that "I am hoping to find a skilled mechanic to work with me in a project to make disconnected cars.")

And later Stallman calls the smartphone "an Orwellian tracking and surveillance device," saying he refuses to own one. (An advantage of free software is that it allows the removal of malicious functionalities.)

Stallman spoke for about 53 minutes — but then answered questions for nearly 90 minutes longer. Here's some of the highlights...

Newsweek reports on how the U.S. Congress is debating "kill switch" technology for vehicles, "which would be able to monitor diver behavior, detect impairment such as intoxication and intervene..."

"While the technology is not yet a legal requirement in cars, Congress passed a law with the Infrastructure Investment and Jobs Act in 2021 that requires the Department of Transportation to create the mandate." Republican Representative Thomas Massie of Kentucky introduced an amendment to a federal spending bill that would reverse the mandating of the technology. On Thursday, 160 Republicans voted in favor, but the legislation failed 164-268, according to the House Clerk's official roll call — with 57 Republicans joining 211 Democrats in voting against it...

The House vote signals substantial Republican support for curbing any move toward mandated impaired-driving prevention systems, but not enough to pass such legislation. Critics of the kill switch technology see it as government overreach, while those in favor argue that it could prove to be lifesaving.
Thanks to long-time Slashdot reader SonicSpike for sharing the article.

This month saw results from a yearlong global study of "potential negative risks that generative AI poses to student". The study (by the Brookings Institution's Center for Universal Education) also suggests how to prevent risks and maximize benefits: After interviews, focus groups, and consultations with over 500 students, teachers, parents, education leaders, and technologists across 50 countries, a close review of over 400 studies, and a Delphi panel, we find that at this point in its trajectory, the risks of utilizing generative AI in children's education overshadow its benefits.
"At the top of Brookings' list of risks is the negative effect AI can have on children's cognitive growth," reports NPR — "how they learn new skills and perceive and solve problems." The report describes a kind of doom loop of AI dependence, where students increasingly off-load their own thinking onto the technology, leading to the kind of cognitive decline or atrophy more commonly associated with aging brains... As one student told the researchers, "It's easy. You don't need to (use) your brain." The report offers a surfeit of evidence to suggest that students who use generative AI are already seeing declines in content knowledge, critical thinking and even creativity. And this could have enormous consequences if these young people grow into adults without learning to think critically...

Survey responses revealed deep concern that use of AI, particularly chatbots, "is undermining students' emotional well-being, including their ability to form relationships, recover from setbacks, and maintain mental health," the report says. One of the many problems with kids' overuse of AI is that the technology is inherently sycophantic — it has been designed to reinforce users' beliefs... Winthrop offers an example of a child interacting with a chatbot, "complaining about your parents and saying, 'They want me to wash the dishes — this is so annoying. I hate my parents.' The chatbot will likely say, 'You're right. You're misunderstood. I'm so sorry. I understand you.' Versus a friend who would say, 'Dude, I wash the dishes all the time in my house. I don't know what you're complaining about. That's normal.' That right there is the problem."
AI did have some advantages, the article points out: The report says another benefit of AI is that it allows teachers to automate some tasks: "generating parent emails ... translating materials, creating worksheets, rubrics, quizzes, and lesson plans" — and more. The report cites multiple research studies that found important time-saving benefits for teachers, including one U.S. study that found that teachers who use AI save an average of nearly six hours a week and about six weeks over the course of a full school year...

AI can also help make classrooms more accessible for students with a wide range of learning disabilities, including dyslexia. But "AI can massively increase existing divides" too, [warns Rebecca Winthrop, one of the report's authors and a senior fellow at Brookings]. That's because the free AI tools that are most accessible to students and schools can also be the least reliable and least factually accurate... "[T]his is the first time in ed-tech history that schools will have to pay more for more accurate information. And that really hurts schools without a lot of resources."
The report calls for more research — and make several recommendations (including "holistic" learning and "AI tools that teach, not tell.") But this may be their most important recommendation. "Provide a clear vision for ethical AI use that centers human agency..."

"We find that AI has the potential to benefit or hinder students, depending on how it is used."

An anonymous reader shared this report from the Guardian: Google's search feature AI Overviews cites YouTube more than any medical website when answering queries about health conditions, according to research that raises fresh questions about a tool seen by 2 billion people each month.

The company has said its AI summaries, which appear at the top of search results and use generative AI to answer questions from users, are "reliable" and cite reputable medical sources such as the Centers for Disease Control and Prevention and the Mayo Clinic. However, a study that analysed responses to more than 50,000 health queries, captured using Google searches from Berlin, found the top cited source was YouTube. The video-sharing platform is the world's second most visited website, after Google itself, and is owned by Google. Researchers at SE Ranking, a search engine optimisation platform, found YouTube made up 4.43% of all AI Overview citations. No hospital network, government health portal, medical association or academic institution came close to that number, they said. "This matters because YouTube is not a medical publisher," the researchers wrote. "It is a general-purpose video platform...."

In one case that experts said was "dangerous" and "alarming", Google provided bogus information about crucial liver function tests that could have left people with serious liver disease wrongly thinking they were healthy. The company later removed AI Overviews for some but not all medical searches... Hannah van Kolfschooten, a researcher specialising in AI, health and law at the University of Basel who was not involved with the research, said: "This study provides empirical evidence that the risks posed by AI Overviews for health are structural, not anecdotal. It becomes difficult for Google to argue that misleading or harmful health outputs are rare cases.

"Instead, the findings show that these risks are embedded in the way AI Overviews are designed. In particular, the heavy reliance on YouTube rather than on public health authorities or medical institutions suggests that visibility and popularity, rather than medical reliability, is the central driver for health knowledge."

"Airlines canceled more than 10,000 U.S. flights scheduled for this weekend," reports CNBC, "as a massive winter storm sweeps across the country, with heavy snow and sleet forecast, followed by bitter cold... set to snarl travel for hundreds of thousands of people for days." More than 3,500 flights on Saturday were canceled, according to flight tracker FlightAware. Many of Saturday's cancellations were in and out of Dallas Fort Worth International Airport, with about 1,300 scrubbed flights, and at Dallas Love Field, with 186 cancellations, the majority of the schedule at each airport. American Airlines, based in Fort Worth, Texas, had canceled 902 Saturday flights, about 30% of its mainline schedule and Southwest Airlines canceled 571 flights, or 19%, according to FlightAware.

U.S. flight cancellations nearly doubled to more than 7,000 [now up to 8,947] on Sunday when the storm is expected to hit the mid-Atlantic and Northeast U.S. As of midday on Saturday, most flights from Raleigh-Durham International Airport in North Carolina to Portland, Maine, were canceled. Major airline hubs were affected as far south as Atlanta, where Delta Air Lines is based.... American, Delta, JetBlue Airways, Southwest Airlines, United Airlines and other U.S. carriers said they are waiving change and cancellation fees as well as fare differences to rebook for customers with tickets to and from more than 40 airports around the country. The waivers include restrictive basic economy tickets.
More than 80% of Sunday's flights at New York's LaGuardian Airport were cancelled, according to the article, at well as 90% of Sunday's flights at Viriginia's Ronald Reagan Washington National Airport.

Slashdot reader Mr. Dollar Ton summarizes this article from Bloomberg: According to a new report from think tank "Ember", the availability of cheap green tech can have developing countries profit from earlier investment and skip steps in the transition from fossil to alternatives.

India is put forward as an example. While China's rapid electrification has been hailed as a miracle, by some measures, India is moving ahead faster than China did when it was at similar levels of economic development. It's an indication that clean electricity could be the most direct way to boost growth for other developing economies.

That's mainly because India has access to solar panels and electric cars at a much lower price than China did about a decade ago. Chinese investments lowered the costs of what experts call "modular technologies" — the production of each solar panel, battery cell and electric car enables engineers to learn how to make it more efficiently.
The think tank's team even argues "that countries such as India, which don't have significant domestic fossil-fuel reserves, will become 'electrostates' that meet most of their energy needs through electricity generated from clean sources," according to the article: No country is an electrostate yet, [says Ember strategist Kingsmill Bond], but countries are increasingly turning to green electricity to power their economies. Nations that are less developed than India will see even more advantages as the cost of electricity technologies, from solar panels and electric vehicles to battery components and minerals, continue to fall.

Neither India nor China is going electric purely to cut emissions or meet climate targets, says Bond. They're doing so because it makes economic sense, particularly for India, which imports more than 40% of its primary energy in the form of coal, oil and gas, according to the International Energy Agency. "To grow and have energy independence, India needs to reduce the terrible burden of fossil-fuel imports worth $150 billion each year," said Bond. "India needs to find other solutions...."

[I]f countries like India find ways to grow electrotech manufacturing without absolute dependence on Chinese equipment, electrification could speed up further. With the U.S. and Europe continuing to add exclusions for Chinese-linked electrotech, countries like India will have an incentive to invest in their own manufacturing capacity. "We are probably at a moment of peak Chinese dominance in the electrotech system, as the rest of the world starts to wake up and realize that this is the energy future," he said.

California became the first U.S. state to join the World Health Organization's Global Outbreak Alert and Response Network (GOARN), one day after the U.S. formally exited the WHO. The Hill reports: This announcement comes just one day after the U.S.'s withdrawal from the WHO became official after nearly 80 years of membership, having been a founding member of the organization. "The Trump administration's withdrawal from WHO is a reckless decision that will hurt all Californians and Americans," [California Governor Gavin Newsom] said in a statement. "California will not bear witness to the chaos this decision will bring. We will continue to foster partnerships across the globe and remain at the forefront of public health preparedness, including through our membership as the only state in WHO's Global Outbreak Alert & Response Network."

Longtime Slashdot reader AmiMoJo shares a report from the Guardian: The financial campaigner James Daley has launched a 1.5 billion pound (approximately $1.5 billion) class action lawsuit against Apple over its mobile phone wallet, claiming the U.S. tech company blocked competition and charged hidden fees that ultimately harmed 50 million UK consumers. The lawsuit takes aim at Apple Pay, which they say has been the only contactless payment service available for iPhone users in Britain over the past decade.

Daley, who is the founder of the advocacy group Fairer Finance, claims this situation amounted to anti-competitive behavior and allowed Apple to charge hidden fees, ultimately pushing up costs for banks that passed charges on to consumers, regardless of whether they owned an iPhone. It is the first UK legal challenge to the company's conduct in relation to Apple Pay, and takes place months after regulators like the Competition and Markets Authority and the Payments Systems Regulator began scrutinising the tech industry's digital wallet services. The case has been filed with the Competition Appeal Tribunal, which will now decide whether the class action case can move forward.

[...] Daley's lawsuit alleges that Apple refused to give other app developers and outside businesses access to the contactless payment technology on its iPhones, which meant it could charge banks and card issuers fees on Apple Pay transactions that his lawyers say "are not in line with industry practice." The lawsuit notes that similar fees are not charged on equivalent payments on Android devices, which are built by Google. It says that the additional costs were borne by UK consumers, having been passed on through charges on a range of personal banking products ranging from current accounts, credit cards, to savings and mortgages. The lawsuit says that about 98% of consumers are exposed to banks that listed cards on Apple Pay, meaning the vast majority of the UK population may have been affected.

An anonymous reader shares a report: A niche retirement plan favored by freelancers is quickly becoming a hot Wall Street sales pitch, as more and more Americans look for ways to shelter a bigger chunk of their paychecks from taxes. Known as solo 401(k)s, they allow the self-employed to contribute $72,000 a year into tax-advantaged retirement accounts. That's nearly three times the maximum for typical salaried workers in the US.

While they've existed for decades serving a workforce that often struggled to earn enough to max out those contributions, wealth planners like JPMorgan Chase & Co. and Betterment are now racing to tap into burgeoning demand from a newer, and wealthier cohort: Post-pandemic contractors and self-employed DIY savers looking to shelter more income, grow assets tax-deferred or tax-free, all with the click of a button.

The pitch is simple: Because of a quirk in the tax code, self-employed workers effectively contribute twice to their 401(k)s -- once as an employee on their own behalf and then again as a business owner making matching contributions. The platforms take care of the paperwork and clients get institutional-level tax planning and investment flexibility. More than three-quarters of America's record 36 million small businesses now have just a single employee, the owner. Cerulli Associates projects that total 401(k) plans in the U.S. will surpass 1 million by 2030, and the fastest growth is expected in sub-$5 million "micro" accounts.

A Toronto man posed as a pilot for years in order to fool airlines into giving him hundreds of free flights, prosecutors have alleged, in a case that has prompted comparisons to the Hollywood thriller Catch Me If You Can. From a report: Authorities in Hawaii announced this week that Dallas Pokornik, 33, had been charged with wire fraud after he allegedly fooled three major US carriers into giving him free tickets over a span of four years.

Airlines typically offer standby tickets to their own staff and those with rival airlines as a way of ensuring the broader industry can effectively move employees across continents. According to court documents, Pokornik was a flight attendant for a Toronto-based airline from 2017 to 2019, but then used an employee identification from that carrier to obtain tickets, "which he in fact knew to be fraudulent at the time it was so presented."

The only Toronto-based airline, Porter, told reporters it was "unable to verify any information related to this story." On one occasion, Pokornik is alleged to have requested a jumpseat in an aircraft's cockpit, which are normally reserved for off-duty pilots, even though he was not a pilot and did not have an airman's certificate. Federal rules prohibit the cockpit jumpseats from being used for leisure travel.

The United States formally withdrew from the World Health Organization on Thursday, making good on an executive order that President Trump issued on his first day in office pledging to leave the international organization that coordinates global responses to public health threats. The New York Times: While the United States is walking away from the organization, a senior official with the Department of Health and Human Services told reporters on Thursday that the Trump administration was considering some type of narrow, limited engagement with W.H.O. global networks that track infectious diseases, including influenza.

As a W.H.O. member, the United States long sent scientists from the Centers for Disease Control and Prevention to participate in international decision-making about which strains to include in the flu vaccine. A W.H.O. meeting on next year's vaccine is scheduled for February. The official said the Trump administration would soon disclose how or whether it will participate.

On Thursday, the administration said that all U.S. government funding to the organization had been terminated, and that all assigned federal employees and contractors had been recalled from its Geneva headquarters and its offices worldwide.

An anonymous reader quotes a report from NPR: TikTok has finalized a deal to create a new American entity, avoiding the looming threat of a ban in the United States that has been in discussion for years. The social video platform company signed agreements with major investors including Oracle, Silver Lake and MGX to form the new TikTok U.S. joint venture. The new version will operate under "defined safeguards that protect national security through comprehensive data protections, algorithm security, content moderation and software assurances for U.S. users," the company said in a statement Thursday. American TikTok users can continue using the same app. [...] Adam Presser, who previously worked as TikTok's head of operations and trust and safety, will lead the new venture as its CEO. He will work alongside a seven-member, majority-American board of directors that includes TikTok's CEO Shou Chew.

[...] In addition to an emphasis on data protection, with U.S. user data being stored locally in a system run by Oracle, the joint venture will also focus on TikTok's algorithm. The content recommendation formula, which feeds users specific videos tailored to their preferences and interests, will be retrained, tested and updated on U.S. user data, the company said in its announcement. The algorithm has been a central issue in the security debate over TikTok. China previously maintained the algorithm must remain under Chinese control by law. But the U.S. regulation passed with bipartisan support said any divestment of TikTok must mean the platform cuts ties -- specifically the algorithm -- with ByteDance. Under the terms of this deal, ByteDance would license the algorithm to the U.S. entity for retraining.

The law prohibits "any cooperation with respect to the operation of a content recommendation algorithm" between ByteDance and a new potential American ownership group, so it is unclear how ByteDance's continued involvement in this arrangement will play out. Oracle, Silver Lake and the Emirati investment firm MGX are the three managing investors, who each hold a 15% share. Other investors include the investment firm of Michael Dell, the billionaire founder of Dell Technologies. ByteDance retains 19.9% of the joint venture.

An anonymous reader quotes a report from Ars Technica: It's no secret that students worldwide use AI chatbots to do their homework and avoid learning things. On the flip side, students can also use AI as a tool to beef up their knowledge and plan for the future with flashcards or study guides. Google hopes its latest Gemini feature will help with the latter. The company has announced that Gemini can now create free SAT practice tests and coach students to help them get higher scores. As a standardized test, the content of the SAT follows a predictable pattern. So there's no need to use a lengthy, personalized prompt to get Gemini going. Just say something like, "I want to take a practice SAT test," and the chatbot will generate one complete with clickable buttons, graphs, and score analysis.

Of course, generative AI can go off the rails and provide incorrect information, which is a problem when you're trying to learn things. However, Google says it has worked with education firms like The Princeton Review to ensure the AI-generated tests resemble what students will see in the real deal. The interface for Gemini's practice tests includes scoring and the ability to review previous answers. If you are unclear on why a particular answer is right or wrong, the questions have an "Explain answer" button right at the bottom. After you finish the practice exam, the custom interface (which looks a bit like Gemini's Canvas coding tool) can help you follow up on areas that need improvement. Google says support for the SAT is just the start, "with more tests coming in the future."

Half the world's 100 largest cities are experiencing high levels of water stress, with 38 of these sitting in regions of "extremely high water stress," new analysis and mapping has shown. The Guardian: Water stress means that water withdrawals for public water supply and industry are close to exceeding available supplies, often caused by poor management of water resources exacerbated by climate breakdown. Watershed Investigations and the Guardian mapped cities on to stressed catchments revealing that Beijing, New York, Los Angeles, Rio de Janeiro and Delhi are among those facing extreme stress, while London, Bangkok and Jakarta are classed as being highly stressed.

Separate analysis of NASA satellite data, compiled by scientists at University College London, shows which of the largest 100 cities have been drying or getting wetter over two decades with places such as Chennai, Tehran and Zhengzhou showing strong drying trends and Tokyo, Lagos and Kampala showing strong wetting trends. All 100 cities and their trends can be viewed on a new interactive water security atlas.

An anonymous reader shares a report: An airport made of bamboo? A tower reaching 20 metres high? For many years, bamboo has been mostly known as the favourite food of giant pandas, but a group of engineers say it's time we took it seriously as a building material, too.

This week the Institution of Structural Engineers called for architects to be "bamboo-ready" as they published a manual for designing permanent buildings made of the material, in an effort to encourage low-carbon construction and position bamboo as a proper alternative to steel and concrete.

Bamboo has already been used for a number of boundary-pushing projects around the world. At Terminal 2 of Kempegowda international airport in Bengaluru, India, bamboo tubes make up the ceiling and pillars. The Ninghai bamboo tower in north-east China, which is more than 20 metres tall, is claimed to be the world's first high-rise building made using engineered bamboo.

At the Green School in Bali, a bamboo-made arc serves as the gymnasium and a striking example of how the material is reshaping sustainable architecture. The use of composite bamboo shear walls have proved to be resilient against earthquakes and extreme weather in countries such as Colombia and the Philippines, where sustainable, disaster-resilient housing has been built with locally sourced materials.

Claims that Chinese technology for AI lags the US are a "fairy tale," Arthur Mensch, the chief executive officer of Mistral, said. From a report: "China is not behind the West," Mensch said in an interview on Bloomberg Television at the World Economic Forum in Davos, Switzerland on Thursday. The capabilities of China's open-source technology is "probably stressing the CEOs in the US."

The remarks from the boss of one of Europe's leading AI companies diverge from other tech leaders at Davos, who reassured lawmakers and business chiefs that China is behind the cutting edge by months or years.

An expanded class-action lawsuit filed last Friday alleges that a member of Nvidia's data strategy team directly contacted Anna's Archive -- the sprawling shadow library hosting millions of pirated books -- to explore "including Anna's Archive in pre-training data for our LLMs."

Internal documents cited in the amended complaint show Nvidia sought information about "high-speed access" to the collection, which Anna's Archive charged tens of thousands of dollars for. According to the lawsuit, Anna's Archive warned Nvidia that its library was illegally acquired and maintained, then asked if the company had internal permission to proceed. The pirate library noted it had previously wasted time on other AI companies that couldn't secure approval. Nvidia management allegedly gave "the green light" within a week.

Anna's Archive promised access to roughly 500 terabytes of data, including millions of books normally only accessible through Internet Archive's controlled digital lending system. The lawsuit also alleges Nvidia downloaded books from LibGen, Sci-Hub, and Z-Library.

An anonymous reader quotes a report from Inside Climate News: Just 32 companies accounted for over half of global fossil carbon emissions in 2024, according to a report published Wednesday by the U.K.-based think tank InfluenceMap. That is down from 36 companies responsible for half the global CO2 emissions in 2023, and 38 companies five years ago. The analysis is the latest update to the Carbon Majors database, which tracks the world's largest oil, gas, coal and cement producers and uses production data to calculate the carbon emissions from each entity's production. The database, first developed by researcher Richard Heede and now hosted by InfluenceMap, quantifies current and historical emissions attributable to nearly 180 companies and provides annual updates. It is the only database of its kind tracking corporate-generated carbon emissions dating back to the start of the Industrial Revolution, research that's being used in efforts to hold major polluters accountable for climate harms.

Despite dire warnings from scientists about the consequences of accelerating climate change, fossil fuel production is continuing apace. Last year, fossil fuel CO2 emissions reached a record high, topping 38 billion metric tons. In 2024 these emissions were 37.4 billion metric tons -- up 0.8 percent from 2023 -- and traceable to 166 oil, gas, coal and cement producers, according to the report. Much of the global carbon emissions in 2024 came from state-owned entities, which represented 16 of the top 20 emitters. The five largest emitters overall -- Saudi Arabia's Aramco, Coal India, China's CHN Energy, National Iranian Oil Co. and Russia's Gazprom -- were all state-controlled, and accounted for 18 percent of the total fossil CO2 emissions in 2024.

ExxonMobil, Chevron, Shell, ConocoPhillips and BP -- the top five emitting investor-owned companies -- together were responsible for 5.5 percent of the total emissions in that year. Historically, ExxonMobil and Chevron rank in the top five for fossil carbon emissions generated from 1854 through 2024, accounting for 2.79 percent and 3.08 percent of overall carbon pollution, respectively. According to the analysis, the 178 entities in the database have generated 70 percent of fossil CO2 emissions since the start of the Industrial Revolution, and just 22 entities are responsible for one-third of these emissions. "Each year, global emissions become increasingly concentrated among a shrinking group of high-emitting producers, while overall production continues to grow. Simultaneously, these heavy emitters continue to use lobbying to obstruct a transition that the scientific community has known for decades is essential," said Emmett Connaire, senior analyst at InfluenceMap. The findings of the new analysis, he added, "underscore the growing importance of this kind of rigorous evidence in efforts to determine accountability for climate-related losses."

Ars Technica's Benj Edwards reports: On Saturday, tech entrepreneur Siqi Chen released an open source plugin for Anthropic's Claude Code AI assistant that instructs the AI model to stop writing like an AI model. Called "Humanizer," the simple prompt plugin feeds Claude a list of 24 language and formatting patterns that Wikipedia editors have listed as chatbot giveaways. Chen published the plugin on GitHub, where it has picked up over 1,600 stars as of Monday. "It's really handy that Wikipedia went and collated a detailed list of 'signs of AI writing,'" Chen wrote on X. "So much so that you can just tell your LLM to... not do that."

The source material is a guide from WikiProject AI Cleanup, a group of Wikipedia editors who have been hunting AI-generated articles since late 2023. French Wikipedia editor Ilyas Lebleu founded the project. The volunteers have tagged over 500 articles for review and, in August 2025, published a formal list of the patterns they kept seeing.

Chen's tool is a "skill file" for Claude Code, Anthropic's terminal-based coding assistant, which involves a Markdown-formatted file that adds a list of written instructions (you can see them here) appended to the prompt fed into the large language model (LLM) that powers the assistant. Unlike a normal system prompt, for example, the skill information is formatted in a standardized way that Claude models are fine-tuned to interpret with more precision than a plain system prompt. (Custom skills require a paid Claude subscription with code execution turned on.)

But as with all AI prompts, language models don't always perfectly follow skill files, so does the Humanizer actually work? In our limited testing, Chen's skill file made the AI agent's output sound less precise and more casual, but it could have some drawbacks: it won't improve factuality and might harm coding ability. [...] Even with its drawbacks, it's ironic that one of the web's most referenced rule sets for detecting AI-assisted writing may help some people subvert it.