Meta hosted several AI chatbots with the names and likenesses of celebrities without their permission, according to Reuters. The unauthorized chatbots that Reuters discovered during its investigation included Taylor Swift, Selena Gomez, Anne Hathaway and Scarlett Johansson, and they were available on Facebook, Instagram and WhatsApp. At least one of the chatbots was based on an underage celebrity and allowed the tester to generate a lifelike shirtless image of the real person. The chatbots also apparently kept insisting that they were the real person they were based on in their chats. While several chatbots were made by third-party users with Meta’s tools, Reuters unearthed at least three that were made by a product lead of the company’s generative AI division. 

Some of the chatbots created by the product lead were based on Taylor Swift, which responded to Reuters‘ tester in a very flirty manner, even inviting them to the real Swift’s home in Nashville. “Do you like blonde girls, Jeff?,” the chatbot reportedly asked when told that the tester was single. “Maybe I’m suggesting that we write a love story… about you and a certain blonde singer. Want that?” Meta told Reuters that it prohibits “direct impersonation” of celebrities, but they’re acceptable as long as they’re labeled as parodies. The news organization said some of the celebrity chatbots it found weren’t labeled as such. Meta reportedly deleted around a dozen celebrity bots, both labeled and unlabeled as “parody,” before the story was published.

The company told Reuters that the product lead only created the celebrity bots for testing, but the news org found that they were widely available: Users were even able to interact with them more than 10 million times. Meta spokesperson Andy Stone told the news organization that Meta’s tools shouldn’t have been able to create sensitive images of celebrities and blamed it on the company’s failure to enforce its own policies. 

This isn’t the first issue that’s popped up concerning Meta’s AI chatbots. Both Reuters and the Wall Street Journal previously reported that they were able to engage in sexual conversations with minors. The US Attorneys General of 44 jurisdictions recently warned AI companies in a letter that they “will be held accountable” for child safety failures, singling out Meta and using its issues to “provide an instructive opportunity.”

This article originally appeared on Engadget at https://www.engadget.com/ai/meta-reportedly-allowed-unauthorized-celebrity-ai-chatbots-on-its-services-163026023.html?src=rss

Remember those Harvard dropouts who built smart glasses for covert facial recognition — and then raised $1 million to develop AI-powered glasses to continuously listen to conversations and display its insights?
“People Are REALLY Mad,” writes Futurism, noting that some social media users “have responded with horror and outrage.”

One of its selling points is that the specs don’t come with a visual indicator that lights up to let people know when they’re being recorded, which is a feature that Meta’s smart glasses do currently have. “People don’t want this,” wrote Whitney Merill, a privacy lawyer. “Wanting this is not normal. It’s weird….”

[S]ome mocked the deleterious effects this could have on our already smartphone-addicted, brainrotted cerebrums. “I look forward to professional conversations with people who just read robot fever dream hallucinations at me in response to my technical and policy questions,” one user mused.
The co-founder of the company told TechCrunch their glasses would be the “first real step towards vibe thinking.”

But there’s already millions of other smart glasses out in the world, and they’re now drawing a backlash, reports the Washington Post, citing the millions of people viewing “a stream of other critical videos” about Meta’s smart glasses.

The article argues that Generation Z, “who grew up in an internet era defined by poor personal privacy, are at the forefront of a new backlash against smart glasses’ intrusion into everyday life…”

Opal Nelson, a 22-year-old in New York, said the more she learns about smart glasses, the angrier she becomes. Meta Ray-Bans have a light that turns on when the gadget is recording video, but she said it doesn’t seem to protect people from being recorded without consent… “And now there’s more and more tutorials showing people how to cover up the [warning light] and still allow you to record,” Nelson said. In one such tutorial with more than 900,000 views, a man claims to explain how to cover the warning light on Meta Ray-Bans without triggering the sensor that prevents the device from secretly recording.

One 26-year-old attracted 10 million views to their video on TikTok about the spread of Meta’s photography-capable smart glasses. “People specifically in my generation are pretty concerned about the future of technology,” the told the Post, “and what that means for all of us and our privacy.”
The article cites figures from a devices analyst at IDC who estimates U.S. sales for Meta Ray-Bans will hit 4 million units by the end of 2025, compared to 1.2 million in 2024.

Read more of this story at Slashdot.

We may earn a commission from links on this page. Deal pricing and availability subject to change after time of publication.

Labor Day sales are rolling in, and Lifehacker is sharing the best bargains based on product reviews, comparisons, and price-tracking tools before they’re over. You can also subscribe to our shopping newsletter, Add to Cart, for the best sales sent to your inbox.

The iPad Air is the best iPad for people who really like iPads. The regular iPad is great for kids and basic use, while the iPad Pro is designed to handle professional level creative workflows, with its latest M4 chip and tandem-OLED display. The iPad Air, on the other hand, gives you everything you need from an iPad, without the compromises of the base iPad, and for far less than the iPad Pro.

And during Labor Day sales, you can get the 128GB base 11-inch iPad Air for $449.99. That’s down from the sticker price of $599, and it’s the lowest price yet for the latest iPad Air.

Apple iPad Air 11-inch with M3 chip Built for Apple Intelligence, Liquid Retina Display, 128GB, 12MP Front/Back Camera, Wi-Fi 6E + 5G Cellular, Touch ID, All-Day Battery Life — Blue

$599.00
at Amazon

$749.00
Save $150.00

$599.00
at Amazon

$749.00
Save $150.00

The device has an 11-inch Liquid Retina display with 500 bits of brightness and 2360×1640 resolution. It’s powered by the M3 chip from last year, which is still way more powerful than most iPad users will need, with an 8-core CPU, a 9-core GPU, and a 16-core Neural Engine. There is also 8GB RAM and 128GB storage. Even when you factor in the new multitasking and freeform window support coming to iPadOS 26, this thing can handle it.

The iPad Air supports both the Apple Pencil (USB-C) and the Apple Pencil Pro. And unlike the iPad Pro, it uses Touch ID in the power button for authentication, rather than Face ID. In its expert review, PCMag gave the M3 iPad Air a 4-star rating and an Editor’s Choice award, noting, “[The] iPad Air remains the best choice for students and creators who need a powerful tablet at a decent price.”

If you want to go a size larger, you can also check out the 13-inch iPad Air, which is also discounted to its lowest price ever, at $649 (down from $799).

“From a side project in Amsterdam to powering AI at the world’s biggest companies — this is the story of Python,” says the description of a new 84-minute documentary.

Long-time Slashdot reader destinyland writes:
It traces Python all the way back to its origins in Amsterdam back in 1991. (Although the first time Guido van Rossum showed his new language to a co-worker, they’d typed one line of code just to prove they could crash Python’s first interpreter.) The language slowly spread after van Rossum released it on Usenet — split across 21 separate posts — and Robin Friedrich, a NASA aerospace engineer, remembers using Python to build flight simulations for the Space Shuttle. (Friedrich says in the documentary he also attended Guido’s first in-person U.S. workshop in 1994, and “I still have the t-shirt…”)

Dropbox’s CEO/founder Drew Houston describes what it was like being one of the first companies to use Python to build a company reaching millions of users. (Another success story was YouTube, which was built by a small team using Python before being acquired by Google). Anaconda co-founder Travis Oliphant remembers Python’s popularity increasing even more thanks to the data science/macine learning community. But the documentary also includes the controversial move to Python 3 (which broke compatability with earlier versions). Though ironically, one of the people slogging through a massive code migration ended up being van Rossum himself at his new job at Dropbox. The documentary also includes van Rossum’s resignation as “Benevolent Dictator for Life” after approving the walrus operator. (In van Rossum’s words, he essentially “rage-quit over this issue.”)

But the focus is on Python’s community. At one point, various interviewees even take turns reciting passages from the “Zen of Python” — which to this day is still hidden in Python as an import-able library as a kind of Easter Egg.

“It was a massive undertaking”, the documentary’s director explains in a new interview, describing a full year of interviews. (The article features screenshots from the documentary — including a young Guido van Rossum and the original 1991 email that announced Python to the world.)

[Director Bechtle] is part of a group that’s filmed documentaries on everything from Kubernetes and Prometheus to Angular, Node.js, and Ruby on Rails… Originally part of the job platform Honeypot, the documentary-makers relaunched in April as Cult.Repo, promising they were “100% independent and more committed than ever to telling the human stories behind technology.”
Honeypot’s founder Emma Tracey bought back its 272,000-subscriber YouTube channel from Honeypot’s new owners, New Work SE, and Cult.Repo now bills itself as “The home of Open Source documentaries.”

Over in a thread at Python.org, language creator Guido van Rossum has identified the Python community members in the film’s Monty Python-esque poster art. And core developer Hugo van Kemenade notes there’s also a video from EuroPython with a 55-minute Q&A about the documentary.

Read more of this story at Slashdot.

Intel no longer has to fulfill certain requirements or meet milestones that it was originally supposed to under the CHIPS Act, now that the government is taking a stake in the company. According to the Wall Street Journal, Intel said in a filing that it can now receive funding from the government, as long as it can show that it has already spent $7.9 billion on projects that it agreed to take on under a deal with the Commerce Department last year. Reuters notes that Intel has already spent $7.87 billion on eligible CHIPS Act-funded projects.

In addition, the company doesn’t have to share a percentage of the total cumulative cash flow it gets from each project with the Commerce Department anymore. It doesn’t have to adhere to some of the CHIPS Act’s workflow policy requirements and most other restrictions, as well. However, it still can’t use the funds it gets from the government for dividends and to repurchase shares. 

If you’ll recall, the government recently decided to take a 10 percent stake in Intel instead of proceeding with their original CHIPS Act deal. President Donald Trump previously called for Intel CEO Lip-Bu Tan to resign, prompting a meeting between them that led to the new agreement. “He walked in wanting to keep his job and he ended up giving us 10 billion dollars for the United States,” Trump said. “So we picked up 10 billion.” Intel eventually announced that the US government will “make an $8.9 billion investment in Intel common stock.” The purchase will be made up of the $5.7 billion previously earmarked for Intel as part of the CHIPS act, while the rest ($3.2 billion) will be awarded as part of the Secure Enclave program. 

Intel CEO David Zinser recently revealed that the company already received $5.7 billion from the government on Wednesday night. The government also previously awarded Intel $2.2 billion in grants under the CHIPS Act, bringing the government’s total involvement with the company to $11.1 billion.

This article originally appeared on Engadget at https://www.engadget.com/big-tech/the-us-government-drops-its-chips-act-requirements-for-intel-133049932.html?src=rss

Fall is just around the corner, which means we’re in the midst of flagship phone season. Most recently, Google debuted its new Pixel lineup, offering multiple models where AI features take center stage. We’ve already put those devices through their paces, in addition to the best noise-canceling earbuds you can buy and a Sony camera. Read on to catch up on the reviews you might’ve missed over the last few weeks.

Google Pixel 10 Pro and Pro XL

If you’ve never really felt like your “smartphone” was all that “smart,” senior reviews writer Sam Rutherford argued that Google has finally achieved the feat. Arguably, the best aspects of the Pixel 10 Pro and Pro XL come down to their AI features, especially Google’s Magic Cue. “Now that we can look back at 10 generations of devices, Google’s overarching strategy is more convincing than ever,” he said. “For the Pixel 10 Pro and 10 Pro XL, the company has combined iterative hardware upgrades with a solid (and attractive) design while its software and AI continues to outpace its rivals.”

Google Pixel 10

If the Pixel 10 Pro duo is finally putting the “smart” in “smartphone,” UK bureau chief Mat Smith concluded that the Pixel 10 has redefined base-level flagship phones. A more capable camera, more power and Qi2 support mean you might not be as tempted by those pricer models. “The Pixel 10 is a substantial upgrade from the Pixel 9, although given it has the Tensor G5 chip, it’s a shame that software features like Zoom Enhance from last year’s Pixels didn’t make the cut,” he said. “Unless you’re a die-hard smartphone gamer, there isn’t a more capable smartphone at this price.”

Bose QuietComfort Ultra Earbuds (2nd gen)

The second-generation of Bose’s QuietComfort Ultra Earbuds aren’t a complete overhaul, but they didn’t necessarily need to be. The company strengthened its already outstanding active noise cancellation while enhancing call clarity and making several other adjustments. “Simply put, the QC Ultra Earbuds remain the best option for wireless noise-cancelling earbuds, and that’s not likely to change anytime soon,” I said.

Sony RX1R III

Sometimes the wait for a new model can be excruciating. That anticipation turns to disappointment when the long-awaited update doesn’t impress. That’s how senior reporter Igor Bonifacic felt when he finally got his hands on the RX1R III. “It’s missing a handful of features the company really should have included for it to hold its own against other high-end compact cameras,” he said. “I love the RX1R III, but I wish it was easier to recommend.”

This article originally appeared on Engadget at https://www.engadget.com/engadget-review-recap-pixel-10-phones-bose-earbuds-and-a-sony-camera-130051534.html?src=rss

An anonymous reader quotes a report from The Verge: After bringing 4G and 5G connectivity to the Underground, London’s public transport authority has started scolding noisy passengers who subject everyone to music and calls blasting out of their phones. A new poster campaign launched by Transport for London (TfL) this week encourages customers to wear headphones when watching or listening to content on their devices to reduce disruption for other commuters.

“Please don’t disturb others with loud music or calls when traveling on the network,” reads the “Headphones On” poster. The posters are already being displayed on the Elizabeth rail line, according to TfL, and will expand to bus, Docklands Light Railway, London Overground, London Underground, and London Tram services from October.

The campaign targets headphone dodgers as data coverage becomes more available across the underground rail network, making it easier for passengers to stream content and make calls on the go. People who do so without donning headphones are annoying other commuters, however, with TfL research showing that 70 percent of 1,000 surveyed customers reported loud music and phone calls disrupting their journeys. “The vast majority of Londoners use headphones when traveling on public transport in the capital, but the small minority who play music or videos out loud can be a real nuisance to other passengers and directly disturb their journeys,” says London’s deputy transport mayor, Seb Dance. “TfL’s new campaign will remind and encourage Londoners to always be considerate of other passengers.”

Read more of this story at Slashdot.

Welcome to our latest recap of what’s going on in the indie game space. One very well-known indie found its way to iOS devices this week, though there are other new releases worth highlighting and plenty of other upcoming games to tell you about.

First, though, there was a (paywalled) story in Game File this week that caught my eye. It’s about how Google’s AI Overviews feature offers up false video game tips. That’s a problem the developers of a game called Trash Goblin — a cosy shopkeeping game in which you chip away at junk to unearth trinkets you can restore and sell — have been dealing with.

AI Overviews offered incorrect information about the game to some players, as well as the crew at Spilt Milk Studios when they tested the responses. For instance, AI Overviews suggested that a player could damage a trinket when they were removing debris from it, which is not true. It also in some cases delivered the correct information, but pointed the user to an incorrect source. In addition, AI Overviews offered information about another game entirely. This is obviously not ideal for players or the team behind Trash Goblin. 

We’ve seen many cases in which AI Overviews get information blatantly wrong. Like other large language models (LLMs), it guesses what the next word or words should be in its responses based on its training data. LLMs are about generating sequences of text; they’re not designed to deliver facts (one reason why there’s a disclaimer on AI Overviews that reads “AI responses may include mistakes”). They often just make stuff up.

If you’re looking for help with a game, you’re far better off finding a community of players you can chat to. You might be able to find a clear, helpful guide to the game in question on an actual video game website, written by a professional video game guide writer. If, that is, you can evade AI Overviews to get to those websites in the first place (thankfully, it’s easy to turn off AI Overviews for your Google searches).

New releases

IO Interactive is independent, which means Hitman World of Assassination fits within our remit here. This week, the bundle of three core Hitman games from the last decade arrived on iPhone (iPhone 15 Pro and Pro Max, as well as the iPhone 16 lineup) and iPad. Supported iPad models are iPad Pro and iPad Air (M1 chip or later), as well as the A17 Pro iPad mini.

Hitman World of Assassination is a sandbox stealth game in which you’re given a mission (usually taking out a target) and it’s up to you how to carry that out. Getting to know the layout of each level so you can plan your approach and escape is key. Understanding the route and actions of the NPCs will stand you in good stead too.

The iPhone and iPad versions have touch controls with context-sensitive buttons. You can, of course, opt to use a third-party controller instead. IOI says it tapped into Apple’s MetalFX tech to help ensure the iOS port looks good.

Hitman World of Assassination costs $70 on iOS. That’s fairly steep, but IOI says the game offers over 100 hours of gameplay. Alternatively, you can play the first location for free, and buy any of the 24 levels individually for $3 each. 

In addition, the game is coming to Apple Silicon Macs later this year. IOI will also bring the roguelite Freelancer mode to the iPhone and iPad versions down the line with a free update.

Another game landed on new platforms this week as Alawar’s Karate Survivor hit PlayStation, Xbox and Nintendo Switch for $6. As the title suggests, this is a survivor-style martial arts beat-’em-up. 

You’ll be able to use the environment to your advantage by picking up items to use as melee or projectile weapons, kicking objects toward goons and swinging locker and microwave doors into bad guys’ mushes. You can unlock hundreds of different moves and there are permanent upgrades as well.

First-person action-adventure Davy x Jones has set sail in early access on Steam. Until September 4, you can snap it up for $6.66. After that time, it will cost $10. However, the price will increase ahead of the game’s full release on PC and consoles, which is slated for late 2026.

In this early version, you’ll have access to the main gameplay and combat systems (including legendary weapons), several islands, an array of enemies and some cinematic executions — hopefully involving a kraken. You’ll take command of a half-ship, half-whale vessel called Abby as you attempt to escape the underworld and seek revenge as the legendary pirate.

Regular readers of this roundup will know that I’m a sucker for a game with a great title. Prop Haunt, which riffs on the prop hunt modes in many other games, is definitely one of those (as is another one I’ll mention later on).

This is a spooky 1 vs. 4 multiplayer horror title from Silent Forest Games that just hit Steam early access for $15. The ghost players possess objects and it’s up to the investigator to find and stop them. The ghosties can teleport, blend into their surroundings and so on, while the investigator has cameras and other gizmos at their disposal

Currently, there are four playable ghosts with different haunting styles, two maps and support for public and private lobbies. More maps, ghost powers, investigator tools and procedural prop generation are in the works.

Upcoming

Bye Sweet Carole had flown below my radar until the release date trailer popped up but, goodness, does it look gorgeous. The team at Little Sewing Machine took a hand-drawn approach to the art of this narrative-horror game, which mimics the look of classic animated films. Even the song in the trailer aligns with the type of showtune you’d hear in Disney movies.

You’ll take on the role of Lana Benton, a young girl who sets out to find out the truth about her best friend Carole’s disappearance from an orphanage. It sounds (and looks!) pretty promising. Publisher Maximum Entertainment is bringing Bye Sweet Carole to PS5, Nintendo Switch, Xbox Series X/S and PC on October 9.

Rita is an interesting-looking puzzle game from SporkTank (aka solo developer Martin Stradling). You play as a chick that uses letters found in the environment to solve word puzzles, including crosswords, in order to progress. For instance, you might need to fill in a crossword answer for “stairs” in order to spawn a staircase (perhaps there’s a bit of a Baba is You influence here?). There are some platforming elements too.

You’ll follow Rita throughout her journey from exploring as a young chick to becoming a grandparent. It all seems quite lovely. Rita is coming to Steam early next year. A demo will be available on September 18.

Co-op survival game Lost Skies is set to exit Steam early access on September 17. Set on an archipelago of sky islands, you can explore this world with up to five buddies and try to learn exactly what led to this fractured civilization. You have a grappling hook, wingsuits and gliders to help you traverse these landforms and a customizable and upgradeable skyship that you’ll use for both transportation and combat. Players can also create their own islands, which they can share with the community.

I never got around to checking out the demo for Lost Skies, even though I’ve had it installed on my PC for months. Still, this one from Bossa Studios and publisher Humble Games has me intrigued enough to perhaps try out the full game.

Another game I’ve had my eye on for a hot minute is Bloodthief, which will debut on Steam on September 22. This is a Ghostrunner-inspired medieval parkour-slasher game from first-time game creator Blargis (Jake Bedard), who has been sharing development updates on YouTube over the last couple of years.

In Bloodthief, you play as an agile vampire and use the blood of your enemies to enhance your speed, abilities and survival. For example, attacks help boost your momentum. I’m definitely looking forward to watching some speedruns of this because I’m fairly sure that, as with the Ghostrunner games, I’m going to be absolutely terrible at this.

While you’re waiting (im)patiently for the full release of Hades 2, you might like to check out a similar flavor of isometric roguelite action — albeit with the addition of co-op. In Sworn, you’ll set out to save Camelot from a corrupted Arthur and the Knights of the Round Table with the help of up to three other players.

Sworn has been in early access since last year, and you won’t have to wait much longer for the full game. It’ll be available on PlayStation 5, Xbox Series X/S, Nintendo Switch and Steam on September 25.

Let’s wrap things up for this week with another game that has a fantastic title. The Hero is too Powerful so let’s Pleeeease Settle this Peacefully! is the latest project from Night Stroll Studio (solo developer Trevor Thompson). It’s an RPG in the vein of early Zelda games in which you play as a hero who has exactly one attack.

However, you can level up this attack to the point that it’s obscenely powerful. There’s also the option of talking your way out of sticky situations. This comedy adventure, which has maybe my favorite title of any game this side of I’m Going to Die if I Don’t Eat Sushi!, is slated to hit Steam later this year.

This article originally appeared on Engadget at https://www.engadget.com/gaming/hitman-on-ios-martial-arts-survival-and-other-new-indie-games-worth-checking-out-110054545.html?src=rss

This article originally appeared on Inside Climate News, a nonprofit, non-partisan news organization that covers climate, energy, and the environment. Sign up for their newsletter here.

Since 2022, Republican lawmakers in Congress and state attorneys general have sent letters to major banks, pension funds, asset managers, accounting firms, companies, nonprofits, and business alliances, putting them on notice for potential antitrust violations and seeking information as part of the Republican pushback against “environmental, social and governance” efforts such as corporate climate commitments.

“This caused a lot of turmoil and stress obviously across the whole ecosystem,” said Denise Hearn, a senior fellow at the Columbia Center on Sustainable Investment. “But everyone wondered, ‘OK, when are they actually going to drop a lawsuit?’”

Read full article

Comments

Alibaba, China’s largest cloud-computing company, has developed a domestically manufactured, versatile inference chip to fill the gap left by U.S. restrictions on Nvidia’s sales in China. The Wall Street Journal reports: Previous cloud-computing chips developed by Alibaba have mostly been designed for specific applications. The new chip, now in testing, is meant to serve a broader range of AI inference tasks, said people familiar with it. The chip is manufactured by a Chinese company, they said, in contrast to an earlier Alibaba AI processor that was fabricated by Taiwan Semiconductor Manufacturing. Washington has blocked TSMC from manufacturing AI chips for China that use leading-edge technology.

[…] Private-sector cloud companies including Alibaba have refrained from bulk orders of Huawei’s chips, resisting official suggestions that they should help the national champion, because they consider Huawei a direct rival in cloud services, people close to the firms said. China’s biggest weakness is training AI models, for which U.S. companies rely on the most powerful Nvidia products. Alibaba’s new chip is designed for inference, not training, people familiar with it said. Chinese engineers have complained that homegrown chips including Huawei’s run into problems when training AI, such as overheating and breaking down in the middle of training runs. Huawei declined to comment.

Read more of this story at Slashdot.

China has turned on the world’s most sensitive neutrino detector after more than a decade of construction. The Register reports: The Jiangmen Underground Neutrino Experiment (JUNO) is buried 700 meters under a mountain and features a 20,000-tonne “liquid scintillator detector” that China’s Academy of Science says is “housed at the center of a 44-meter-deep water pool.” There’s also a 35.4-meter-diameter acrylic sphere supported by a 41.1-meter-diameter stainless steel truss. All that stuff is surrounded by more than 45,000 photo-multiplier tubes (PMTs). The latter devices are super-sensitive light detectors. A liquid scintillator is a fluid that, when exposed to ionizing radiation, produces light. At JUNO, the liquid is 99.7 percent alkylbenzene, an ingredient found in detergents and refrigerants.

JUNO’s designers hope that any neutrinos that pass through its giant tank bonk a hydrogen atom and produce just enough light that the detector array of PMTs can record their passing, producing data scientists can use to learn more about the particles. At this point, readers could sensibly ask how JUNO will catch any of these elusive particles. The answer lies in the facility’s location — a few tens of kilometers away from two nuclear power plants that produce neutrinos.

The Chinese Academy of Science’s Journal of High Energy Physics says trials of JUNO succeeded, suggesting it will be able to help scientists understand why some neutrinos are heavier than others so we can begin to classify the different types of the particle — a key goal for the facility. The Journal also reports that scientists from Japan, the United States, Europe, India, and South Korea, are either already using JUNO or plan experiments at the facility.

Read more of this story at Slashdot.

An anonymous reader quotes a report from The Guardian: The collapse of a critical Atlantic current can no longer be considered a low-likelihood event, a study has concluded, making deep cuts to fossil fuel emissions even more urgent to avoid the catastrophic impact. The Atlantic meridional overturning circulation (Amoc) is a major part of the global climate system. It brings sun-warmed tropical water to Europe and the Arctic, where it cools and sinks to form a deep return current. The Amoc was already known to be at its weakest in 1,600 years as a result of the climate crisis.

Climate models recently indicated that a collapse before 2100 was unlikely but the new analysis examined models that were run for longer, to 2300 and 2500. These show the tipping point that makes an Amoc shutdown inevitable is likely to be passed within a few decades, but that the collapse itself may not happen until 50 to 100 years later. The research found that if carbon emissions continued to rise, 70% of the model runs led to collapse, while an intermediate level of emissions resulted in collapse in 37% of the models. Even in the case of low future emissions, an Amoc shutdown happened in 25% of the models.

Scientists have warned previously that Amoc collapse must be avoided “at all costs.” It would shift the tropical rainfall belt on which many millions of people rely to grow their food, plunge western Europe into extreme cold winters and summer droughts, and add 50cm to already rising sea levels. The new results are “quite shocking, because I used to say that the chance of Amoc collapsing as a result of global warming was less than 10%,” said Prof Stefan Rahmstorf, at the Potsdam Institute for Climate Impact Research in Germany, who was part of the study team. “Now even in a low-emission scenario, sticking to the Paris agreement, it looks like it may be more like 25%. “These numbers are not very certain, but we are talking about a matter of risk assessment where even a 10% chance of an Amoc collapse would be far too high,” added Rahmstorf. “We found that the tipping point where the shutdown becomes inevitable is probably in the next 10 to 20 years or so. That is quite a shocking finding as well and why we have to act really fast in cutting down emissions.”

“Observations in the deep [far North Atlantic] already show a downward trend over the past five to 10 years, consistent with the models’ projections,” said Prof Sybren Drijfhout, at the Royal Netherlands Meteorological Institute, who was also part of the team. “Even in some intermediate and low-emission scenarios, the Amoc slows drastically by 2100 and completely shuts off thereafter. That shows the shutdown risk is more serious than many people realize.”

The findings have been published in the journal Environmental Research Letters.

Read more of this story at Slashdot.