ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

OpenAI Says It Won't Release a Model Called Orion This Year (techcrunch.com)

(Saturday October 26, 2024 @03:00AM (msmash) from the for-the-record dept.)

An anonymous reader shares a report:

> OpenAI says that it [1]doesn't intend to release an AI model code-named Orion this year, countering recent reporting on the company's product roadmap. "We don't have plans to release a model code-named Orion this year," a spokesperson told TechCrunch via email. "We do plan to release a lot of other great technology."

>

> The Verge reported on Thursday that Orion, which is expected to be OpenAI's next frontier model, would launch by December, and that trusted partners would be the first to preview it ahead of a rollout through ChatGPT. According to The Verge, Microsoft, a close OpenAI collaborator and investor, expects to gain access to Orion as early as November.



[1] https://techcrunch.com/2024/10/25/openai-says-it-wont-release-a-model-called-orion-this-year/



PayPal To Share Customer Purchase Data with Retailers (msn.com)

(Saturday October 26, 2024 @03:00AM (msmash) from the up-next dept.)

PayPal will begin sharing detailed customer purchase data, including clothing sizes and shopping preferences, with [1]retailers for targeted advertising starting November 27 , the payments company announced in a recent privacy update. The initiative affects PayPal's 391 million active consumer accounts worldwide. While customers can opt out through the app's settings, the GAO reports such opt-out rates typically remain below 7% across financial services.



[1] https://www.msn.com/en-us/money/other/it-has-your-money-and-your-pants-size-here-s-what-paypal-is-doing-with-it/ar-AA1sUjKs



San Francisco Billboards Call Out Tech Firms For Not Paying For Open Source (theregister.com)

(Friday October 25, 2024 @11:30PM (msmash) from the name-and-shame dept.)

An anonymous reader shares a report:

> Drivers passing through San Francisco have a new roadside distraction to consider: [1]billboards calling out businesses that don't cough up for the open source code that they use. The signs are the work of the Open Source Pledge -- a group that launched earlier this month. It asks businesses that make use of open source code to pledge $2,000 per developer to support projects that develop the code. So far, 25 companies have signed up -- but project co-founder Chad Whitacre wants bigger firms to pay their dues, too.

>

> Whitacre, whose day job is head of open source at app-monitoring biz Sentry, told The Register his employer has for three years operated a scheme to pay developers who maintain and upgrade open source code. "We do dollars per developer, the thinking being it's the developers and software engineers on the staff at a company who benefit the most from open source, who become more productive because of open source," he said. "I had one conversation with a representative from a larger firm and he's like: 'Chad, you're asking me to spend ten million on maintainers.'" Whitacre affirmed that request, and pointed out the firm "spends ten million on something anyway."



[1] https://www.theregister.com/2024/10/25/open_source_funding_ads/



Study Finds UBI Results Are Not Positive (nber.org)

(Friday October 25, 2024 @05:20PM (msmash) from the closer-look dept.)

[1]Seven Spirals writes:

> A [2]working paper [PDF] , [3]published by the National Bureau of Economic Research , studies the employment effects of a guaranteed income by providing $1,000 per month to 1,000 low-income participants for three years, compared to a control group receiving $50 per month. The results show a decrease in labor market participation by 2 percentage points and a reduction of 1.3-1.4 hours in weekly work hours. Most of the additional free time was spent on leisure, and there were no significant improvements in job quality or human capital investments. Overall, the guaranteed income led to a moderate reduction in labor supply without other substantial productive benefits.



[1] https://slashdot.org/~Seven+Spirals

[2] https://www.nber.org/system/files/working_papers/w32719/w32719.pdf

[3] https://www.nber.org/papers/w32719



Hacker Returns $19.3 Million To Drained US Government Crypto Wallet (decrypt.co)

(Friday October 25, 2024 @05:20PM (msmash) from the stranger-things dept.)

A government-controlled wallet that had been drained of $20 million on Thursday [1]received most of its funds back Friday, adding another layer of mystery to transactions flagged by blockchain analysts as likely being connected to a high-profile theft. From a report:

> The pseudonymous blockchain sleuth ZachXBT had said in a tweet Thursday that the transfers resembled the playbook of a bad actor. Engaging with several decentralized finance protocols, the wallet had also tapped so-called instant exchanges after funds were moved across a series of transfers that "looked nefarious." About $19.3 million worth of funds had been returned to the wallet early Friday, per on-chain data collected by Arkham Intelligence, including Ethereum and the stablecoin USDC. Still, ZachXBT said in his Telegram community that funds transferred to exchanges had not yet been returned.



[1] https://decrypt.co/288296/hacker-returns-millions-drained-us-government-crypto-wallet



'We Don't Know Where the Tipping Point Is': Climate Expert on Potential Collapse of Atlantic Circulation

(Friday October 25, 2024 @05:20PM (msmash) from the grave-concerns dept.)

Scientists from 15 countries warned in an open letter that the potential collapse of the Atlantic Ocean's main circulation system (AMOC) has been "greatly underestimated" and [1]could have devastating global impacts , according to oceanographer Stefan Rahmstorf who spoke to The Guardian. The AMOC system, which moves heat through the northern Atlantic and influences weather patterns across Europe, has shown signs of weakening over the past 60-70 years due to global warming, with indicators including a persistent cold spot in the northern Atlantic and record-low salt levels in seawater.

Rahmstorf, who heads the Earth system analysis department at Germany's Potsdam Institute for Climate Impact Research, estimates a 50% chance of the system reaching a critical tipping point this century, which could trigger severe consequences including regional cooling in northwestern Europe, shifting rainfall patterns, and rising sea levels.



[1] https://www.theguardian.com/environment/2024/oct/23/we-dont-know-where-the-tipping-point-is-climate-expert-on-potential-collapse-of-atlantic-circulation



Microsoft Bets on Latest 'Call of Duty' To Power Up Video Games Strategy (ft.com)

(Friday October 25, 2024 @05:20PM (msmash) from the big-test dept.)

Microsoft is seeking to boost its video games business with the release of the latest instalment of the Call of Duty franchise on Friday, pushing to increase subscription revenues through the new game to offset falling Xbox console sales. Financial Times:

> Black Ops 6 is the first of the best-selling series to be launched on the tech giant's Game Pass subscription service. It represents [1]the biggest test of the company's gaming strategy [ [2]non-paywalled link ] since its $75bn deal to acquire Activision Blizzard -- makers of Call of Duty -- received sign-off from regulators last year. Microsoft hopes that the release will help achieve its target of reaching 110mn Game Pass subscribers by 2030, a substantial rise from 34mn in February this year.

>

> The company has shifted its focus towards its subscription games service as hardware sales have slowed in recent years. Xbox hardware revenue fell 13 per cent year-on-year in Microsoft's fiscal 2024, which ended in June. For the first time this year, subscribers to Game Pass, who can already access a growing library of Xbox titles for as long as they keep paying a monthly fee, will be able to access the latest Call of Duty without having to pay a traditional price of $70 or more for the packaged game. Microsoft is still making the game available to buy on PlayStation, after concerns from regulators during the Activision merger probe that it might make the title exclusive to its own platform.



[1] https://www.ft.com/content/0b226cad-116d-4125-baf9-fe9295a6e503

[2] https://variety.com/2024/gaming/news/call-of-duty-black-ops-6-activision-xbox-microsoft-1236185153/



SpaceX Brings Home Astronauts After Boeing's Starliner Delays Extend ISS Mission

(Friday October 25, 2024 @05:20PM (msmash) from the getting-job-done dept.)

Four astronauts splashed down in the Gulf of Mexico on Friday after their record ISS mission stretched to eight months due to Boeing capsule malfunctions and hurricane disruptions. The SpaceX Dragon capsule [1]landed off Florida's coast before dawn , carrying NASA astronauts Matthew Dominick, Michael Barratt, Jeanette Epps and Russian cosmonaut Alexander Grebenkin.

Technical issues with Boeing's Starliner capsule in September, followed by Hurricane Milton and persistent rough seas, delayed their planned return by two months. The crew launched in March as part of NASA's commercial crew program. Their replacements include Boeing Starliner test pilots Butch Wilmore and Suni Williams, whose mission expanded from eight days to eight months, alongside two SpaceX-launched astronauts. The new crew will remain aboard the station until February.



[1] https://www.theguardian.com/science/2024/oct/25/spacex-astronauts-return-to-earth-after-delay-due-to-boeing-capsule-snags



Europe In Talks With SpaceX On Tackling Space Junk (reuters.com)

(Friday October 25, 2024 @05:20PM (BeauHD) from the growing-problem dept.)

An anonymous reader quotes a report from Reuters:

> The European Space Agency is in talks with SpaceX about the possibility of Elon Musk's space venture [1]joining an international charter designed to reduce a growing swarm of debris in space , Director General Josef Aschbacher told Reuters. The 22-nation agency is spearheading one of several efforts to roll back the mass of space junk swirling round the planet from past missions that poses a risk to active satellites. Aschbacher said 110 countries or entities have joined ESA's Zero Debris charter, which aims to stop any new orbital garbage being generated by 2030.

>

> Asked whether SpaceX, whose satellites now make up some two thirds of spacecraft active in low Earth orbit, had signed up, Aschbacher said: "Not yet, but we are in discussion with them... This is a charter that keeps evolving and... we will keep raising the topics because they are so fundamental." [...] There are currently 18,897 pieces of trackable space junk in orbit, according to Jonathan McDowell, a Harvard astronomer who tracks such objects. Space debris and junk are often used interchangeably, but some consider space junk to include inactive payloads and rocket bodies as well as debris, or errant shards of broken satellites. There are no international laws on debris, but countries and space agencies have begun in recent years to devise proposals and national rules for tackling the problem.



[1] https://www.reuters.com/technology/space/europe-agency-says-it-is-talks-with-spacex-tackling-space-junk-2024-10-24/



Polish Radio Station Replaces Journalists With AI 'Presenters'

(Friday October 25, 2024 @05:20PM (BeauHD) from the only-the-beginning dept.)

OFF Radio Krakow sparked controversy by [1]replacing its journalists with AI-generated presenters in an experiment to attract younger audiences. CNN Business reports:

> Weeks after letting its journalists go, OFF Radio Krakow relaunched this week, with what it said was Ć¢oethe first experiment in Poland in which journalists ... are virtual characters created by AI." The station in the southern city of Krakow said its three avatars are designed to reach younger listeners by speaking about cultural, art and social issues including the concerns of LGBTQ+ people. "Is artificial intelligence more of an opportunity or a threat to media, radio and journalism? We will seek answers to this question," the station head, Marcin Pulit, wrote in a statement.



[1] https://www.cnn.com/2024/10/24/media/polish-radio-ai-presenters-scli-intl/index.html



UnitedHealth Says Change Healthcare Hack Affects Over 100 Million (techcrunch.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the nothing-to-be-proud-of dept.)

UnitedHealth Group said a ransomware attack in February resulted in [1]more than 100 million individuals having their private health information stolen . The U.S. Department of Health and Human Services [2]first reported the figure on Thursday. TechCrunch reports:

> The ransomware attack and data breach at Change Healthcare stands as the largest known digital theft of U.S. medical records, and one of the biggest data breaches in living history. The ramifications for the millions of Americans whose private medical information was irretrievably stolen are likely to be life lasting. UHG began notifying affected individuals in late July, which continued through October. The stolen data varies by individual, but Change previously confirmed that it includes personal information, such as names and addresses, dates of birth, phone numbers and email addresses, and government identity documents, including Social Security numbers, driver's license numbers, and passport numbers. The stolen health data includes diagnoses, medications, test results, imaging and care and treatment plans, and health insurance information -- as well as financial and banking information found in claims and payment data taken by the criminals.

>

> The cyberattack [3]became public on February 21 when Change Healthcare pulled much of its network offline to contain the intruders, causing immediate outages across the U.S. healthcare sector that relied on Change for handling patient insurance and billing. UHG attributed the cyberattack to ALPHV/BlackCat, a Russian-speaking ransomware and extortion gang, which later took credit for the cyberattack. The ransomware gang's leaders later vanished after absconding with a $22 million ransom paid by the health insurance giant, stiffing the group's contractors who carried out the hacking of Change Healthcare out of their new financial windfall. The contractors took the data they stole from Change Healthcare and formed a new group, which extorted a second ransom from UHG, while publishing a portion of the stolen files online in the process to prove their threat.

>

> There is no evidence that the cybercriminals subsequently deleted the data. Other extortion gangs, including LockBit, have been shown to hoard stolen data, even after the victim pays and the criminals claim to have deleted the data. In paying the ransom, Change obtained a copy of the stolen dataset, allowing the company to identify and notify the affected individuals whose information was found in the data. Efforts by the U.S. government to catch the hackers behind ALPHV/BlackCat, one of the most prolific ransomware gangs today, have so far failed. The gang bounced back following a takedown operation in 2023 to seize the gang's dark web leak site. Months after the Change Healthcare breach, the U.S. State Department upped its reward for information on the whereabouts of the ALPHV/BlackCat cybercriminals to $10 million.



[1] https://techcrunch.com/2024/10/24/unitedhealth-change-healthcare-hacked-millions-health-records-ransomware/

[2] https://www.documentcloud.org/documents/25250169-change-healthcare-breach

[3] https://tech.slashdot.org/story/24/02/22/144230/us-health-tech-giant-change-healthcare-hit-by-cyberattack



Carbon Emissions Are Now Growing Faster Than Before the Pandemic (newscientist.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the not-what-we-want-to-see dept.)

An anonymous reader quotes a report from New Scientist:

> As the world emerged from the lockdowns of the covid-19 pandemic, many countries promised to rebuild their economies in a climate-friendly fashion, amid hopes the recovery effort could accelerate the global journey to net-zero emissions. In reality, the opposite has happened. Instead of a "green recovery," global greenhouse gas emissions are [1]rising much faster now than they did in the decade preceding the global pandemic . Emissions rose 1.3 percent to 57.1 gigatons of carbon dioxide equivalent in 2023. That is a far faster annual rate of growth than during the decade 2010-2019, when emissions grew on average 0.8 percent per year. In fact, global greenhouse gas emissions are now just below the 59.1 gigatons peak recorded in 2019.

>

> All sources of greenhouse gas emissions except land use are rising, according to [2]a report from the United Nations Global Environment Program (UNEP), as economies continued to rebound from covid-19. Emissions from road transport, leaks from oil and gas infrastructure such as pipelines and industrial emissions all grew rapidly in 2023, UNEP said, while aviation emissions grew 19.5 percent. Rising emissions means the world's opportunity to avert catastrophic climate change is shrinking, Inger Anderson at UNEP said in a statement. "Climate crunch time is here," she said. "I urge every nation: no more hot air, please."



[1] https://www.newscientist.com/article/2453198-carbon-emissions-are-now-growing-faster-than-before-the-pandemic/

[2] https://www.unep.org/resources/emissions-gap-report-2024



Google Photos Will Soon Show You If an Image Was Edited With AI

(Friday October 25, 2024 @11:21AM (BeauHD) from the it's-a-start dept.)

Starting next week, Google Photos will [1]label when an image was edited with AI . The Verge reports:

> "Photos edited with tools like Magic Editor, Magic Eraser and Zoom Enhance already include metadata based on technical standards from The International Press Telecommunications Council (IPTC) to indicate that they've been edited using generative AI," John Fisher, engineering director of Google Photos, wrote in a [2]blog post . "Now we're taking it a step further, making this information visible alongside information like the file name, location and backup status in the Photos app."

>

> The "AI info" section will be found in the image details view of Google Photos both on the web and in the app. These labels won't be limited strictly to generative AI, either. Google says it'll also specify when a "photo" contains elements from several different images -- such as when people use the Pixel's Best Take and Add Me features. [...] "This work is not done, and we'll continue gathering feedback and evaluating additional solutions to add more transparency around AI edits," Fisher wrote.



[1] https://www.theverge.com/2024/10/24/24278663/google-photos-generative-ai-label-reimagine-best-take

[2] https://blog.google/products/photos/ai-editing-transparency/



Europe's Top Court Rules For Intel To End Long-Running Antitrust Case (reuters.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the case-closed dept.)

The EU Court of Justice [1]ruled in favor of Intel , dismissing the European Commission's appeal and ending a [2]nearly two-decade-long case over allegations that Intel's rebates to computer makers were anticompetitive. Reuters reports:

> The European Commission had fined Intel for giving rebates to computer makers Dell, Hewlett-Packard and Lenovo for buying most of their chips from Intel, which regulators said was an attempt to block Advanced Micro Devices. Regulators generally oppose rebates offered by dominant companies because they fear they may be anticompetitive, while companies say enforcers must prove discounts have anticompetitive effects before companies are sanctioned.

>

> EU regulators had initially fined Intel 1.06 billion euros ($1.14 billion) but a lower tribunal scrapped that. Intel's case was boosted earlier this year when an adviser to the court said regulators had not properly performed an economic analysis.



[1] https://www.reuters.com/technology/europes-top-court-rules-intel-long-running-antitrust-case-2024-10-24/

[2] https://yro.slashdot.org/story/09/05/10/2357257/sources-say-eu-will-find-intel-anti-competitive



Intel Arrow Lake Core Ultra 200S Tested: Competitively Priced With Creator Performance (hothardware.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the new-and-improved dept.)

[1]MojoKid writes:

> Intel has lifted the embargo on independent reviews of its new Core Ultra 200S series Arrow Lake-S processors, which mark a shift in its desktop CPU strategy with symmetrical core/thread counts (no Hyperthreading) and a dedicated 13 TOPS NPU. This series features a disaggregated tiled design for the first time in Intel's desktop chips, focusing on efficiency and power reduction. The Core Ultra 5 245, priced around $300, and the Ultra 9 285K at $589 [2]deliver strong performance, particularly in creator workloads , competing well with AMD's Ryzen CPUs, while the Ultra 9 285K's price undercuts AMD's Ryzen 9 9950X significantly. While gaming performance shows slight regression in spots, the new chips are much more power-efficient than their predecessors. Overall, the platform offers leading-edge features, competitive pricing, and solid performance for creators, gamers and workstation pros.



[1] https://slashdot.org/~MojoKid

[2] https://hothardware.com/reviews/intel-core-ultra-200s-arrow-lake-cpu-review



Goodreads' Founder Debuts AI-Powered App For Online Readers (techcrunch.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the smashing-success dept.)

An anonymous reader quotes a report from TechCrunch:

> [1]Smashing , a new app curating the best of the web from Goodreads co-founder Otis Chandler, is [2]now available to the public . Like Goodreads, the app aims to create a community around content. But this time, instead of books, the focus is on web content -- like news articles, blog posts, social media posts, podcasts, and more. In addition, Smashing is introducing an AI Questions feature that allows you to engage with the content being shared in different ways, including by viewing a news story from different perspectives or asking the AI to poke holes in the story, among other things. By viewing different angles of a story, you can see how both the political left and right view the subject. Or, in the case of a company's stock, you might be presented with both the bull and bear case.

>

> There are a good handful of AI prompts available at launch, notes Chandler, and not all will make sense to use on every news story or piece of content. For instance, there's a silly "make it funny" prompt, and others that can simplify the story, display a timeline, or introduce "unconventional" takes that may involve thinking outside the box, helping you weigh ideas you hadn't considered yet. You can also ask your own questions, if you prefer. On the app, users are able to create multiple interest feeds to stay informed about the topics that matter to them, like politics, investing, parenting, health and wellness, and more, or even narrower interests like specific companies, sports teams, crypto, climate change, or other subtopics. The app also leverages AI to surface content from around the web and then match it to an individual reader based on what articles they tend to read, what subtopics they like, and what's already popular in the community, as determined by upvotes and downvotes. Combined, the signals tune Smashing to a user's particular interests. As part of the AI Questions feature, Smashing is also introducing AI-powered Story Overview pages, which offer grouped articles, blog posts, and social media posts all about the same story.



[1] https://www.smashing.xyz/

[2] https://techcrunch.com/2024/10/24/smashing-an-ai-powered-app-for-online-readers-launches-to-the-public/



Georgian Authorities Raid Homes of Disinformation Researchers Ahead of Elections (therecord.media)

(Friday October 25, 2024 @11:21AM (BeauHD) from the cease-and-desist dept.)

Ahead of Georgia's parliamentary elections, Georgian authorities [1]raided the homes of disinformation researchers Eto Buziashvili and Sopo Gelava , seizing personal devices. The Record:

> Eto Buziashvili and Sopo Gelava, both employees of the Atlantic Council think tank, had their homes searched and their own and their family members' personal devices seized by investigators working for the country's Ministry of Finance, according to friends of the pair who spoke to Recorded Future News. Both women are said to be safe, although there are concerns about the security of their devices and online accounts. The searches come a day after Buziashvili published [2]an article detailing how the Kremlin was influencing Georgian politics by supporting the incumbent government and interfering in the upcoming elections.

>

> Local media reported that the offices of outsourcing company Concentrix and other Georgian citizens were also subject to searches. The Ministry of Finance [3]claimed on Facebook it launched searches of "specific facilities" related to "call centers" alleged to be engaged in illegal activity. The investigations come ahead of an election that is being seen as a bellwether of the country's future direction, either pursuing closer ties to Russia under the current prime minister Irakli Kobakhidze or moving towards the West through opposition figures.

Graham Brookie, the Atlantic Council's vice president for technology programs and strategy, said the organization "is deeply concerned about this development and its impact on our staff's work shortly before Georgian elections. [Gelava and Buziashvili] are engaged in independent, non-partisan work aimed at defending and strengthening democracy from those who would undermine it in online spaces, including research related to foreign influence efforts, the targeting of marginalized communities, and other online harms."

"We trust that Georgian authorities will provide more clarity on their actions, ensure the safety and security of our staff, return their property, and allow them to continue their contributions to Georgian democracy."



[1] https://therecord.media/georgian-authorities-raid-homes-disinformation-researchers-election

[2] https://www.atlanticcouncil.org/blogs/new-atlanticist/russia-is-directly-and-indirectly-meddling-in-georgias-upcoming-election/

[3] https://www.facebook.com/sagamodziebo/posts/pfbid0t8HUoSe5jgAHQuzmTWc6jhfW6zvr61ALUUSVtBG6pge7HDmb5Xq78mHkbzhhB1Wel



Amazon To Shut Down Speedy Brick-and-Mortar Delivery Service (theverge.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the cost-cutting-efforts dept.)

Amazon is [1]shutting down its Amazon Today service , which provided same-day delivery from mall and brick-and-mortar retailers. The program will mostly wind down by December 2, 2024, with select partners fulfilling orders until January 2025. CNBC reports:

> A small amount of employees will be laid off and provided with severance, while others will be transitioned to other positions within Amazon, the company said. Employees who work on Amazon Today learned the news in a meeting on Monday, where some staffers were informed they would be laid off, the people said. Roughly 300 employees were working on Amazon Today, the people said. Amazon disputed the figure, saying there were approximately 175 employees in the unit. The closure of Amazon Today is the latest example of the company's broader cost-cutting efforts.



[1] https://www.theverge.com/2024/10/24/24278663/google-photos-generative-ai-label-reimagine-best-take



White House Orders Pentagon and Intel Agencies To Increase Use of AI (msn.com)

(Friday October 25, 2024 @03:00AM (msmash) from the no-time-for-caution dept.)

The White House is directing the Pentagon and intelligence agencies to [1]increase their adoption of AI , expanding the Biden administration's efforts to curb technological competition from China and other adversaries. From a report:

> The edict is part of a landmark national security memorandum published Thursday. It aims to make government agencies step up experiments and deployments of AI. The memo also bans agencies from using the technology in ways that "do not align with democratic values," according to a White House news release.

>

> "This is our nation's first ever strategy for harnessing the power and managing the risks of AI to advance our national security," national security adviser Jake Sullivan said in a speech Thursday. Sullivan called the speed of change in AI "breathtaking" and said it had the potential to affect fields ranging from nuclear physics to rocketry and stealth technology. The White House believes that providing clear rules for using AI will make it easier for government agencies to use the technology, according to a briefing with senior administration officials who spoke on the condition of anonymity to discuss details of the report before its publication.



[1] https://www.msn.com/en-us/news/politics/white-house-orders-pentagon-and-intel-agencies-to-increase-use-of-ai/ar-AA1sQb9d



Google Offers Its AI Watermarking Tech As Free Open Source Toolkit (arstechnica.com)

(Friday October 25, 2024 @11:21AM (BeauHD) from the open-source-FTW dept.)

An anonymous reader quotes a report from Ars Technica:

> Back in May, Google [1]augmented its Gemini AI model with SynthID , a toolkit that embeds AI-generated content with watermarks it says are "imperceptible to humans" but can be easily and reliably detected via an algorithm. Today, Google [2]took that SynthID system open source , offering the same basic watermarking toolkit for [3]free to developers and businesses . The [4]move gives the entire AI industry an easy, seemingly robust way to silently mark content as artificially generated, which could be useful for detecting deepfakes and other damaging AI content before it goes out in the wild. But there are still some important limitations that may prevent AI watermarking from becoming a de facto standard across the AI industry any time soon.

>

> Google uses a version of SynthID to watermark audio, video, and images generated by its multimodal AI systems, with differing techniques that are [5]explained briefly in this video . But in a new paper [6]published in Nature , Google researchers go into detail on how the SynthID process embeds an unseen watermark in the text-based output of its Gemini model. The core of the text watermarking process is a sampling algorithm inserted into an LLM's usual token-generation loop (the loop picks the next word in a sequence based on the model's complex set of weighted links to the words that came before it). Using a random seed generated from a key provided by Google, that sampling algorithm increases the correlational likelihood that certain tokens will be chosen in the generative process. A scoring function can then measure that average correlation across any text to determine the likelihood that the text was generated by the watermarked LLM (a threshold value can be used to give a binary yes/no answer).



[1] https://deepmind.google/discover/blog/watermarking-ai-generated-text-and-video-with-synthid/

[2] https://arstechnica.com/ai/2024/10/google-offers-its-ai-watermarking-tech-as-free-open-source-toolkit/

[3] https://x.com/GoogleDeepMind/status/1849110263871529114

[4] https://ai.google.dev/responsible/docs/safeguards/synthid

[5] https://www.youtube.com/supported_browsers?next_url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D9btDaOcfIMY

[6] https://www.nature.com/articles/s41586-024-08025-4?error=cookies_not_supported&code=2a7670ef-5e32-4e42-a266-d9db9bf2c83d



More

... The Anarchists' [national] anthem is an international anthem that
consists of 365 raspberries blown in very quick succession to the tune
of "Camptown Races". Nobody has to stand up for it, nobody has to
listen to it, and, even better, nobody has to play it.
-- Mike Harding, "The Armchair Anarchist's Almanac"