French private AI lab PleIAs "is committed to training LLMs in the open," they write in a blog post at Mozilla.org. "This means not only releasing our models but also being open about every aspect, from the training data to the training code. We define 'open' strictly: all data must be both accessible and under permissive licenses."

Wednesday PleIAs announced they were releasing the largest open multilingual pretraining dataset, according to their blog post at HuggingFace: Many have claimed that training large language models requires copyrighted data, making truly open AI development impossible. Today, Pleias is proving otherwise with the release of Common Corpus (part of the AI Alliance Open Trusted Data Initiative) — the largest fully open multilingual dataset for training LLMs, containing over 2 trillion tokens of permissibly licensed content with provenance information (2,003,039,184,047 tokens).

As developers are responding to pressures from new regulations like the EU AI Act, Common Corpus goes beyond compliance by making our entire permissibly licensed dataset freely available on HuggingFace, with detailed documentation of every data source. We have taken extensive steps to ensure that the dataset is high-quality and is curated to train powerful models. Through this release, we are demonstrating that there doesn't have to be such a [heavy] trade-off between openness and performance.

Common Corpus is:

— Truly Open: contains only data that is permissively licensed and provenance is documented

— Multilingual: mostly representing English and French data, but contains at least 1B tokens for over 30 languages

— Diverse: consisting of scientific articles, government and legal documents, code, and cultural heritage data, including books and newspapers

— Extensively Curated: spelling and formatting has been corrected from digitized texts, harmful and toxic content has been removed, and content with low educational content has also been removed.


Common corpus builds on a growing ecosystem of large, open datasets, such as Dolma, FineWeb, RefinedWeb. The Common Pile currently in preparation under the coordination of Eleuther is built around the same principle of using permissible content in English language and, unsurprisingly, there were many opportunities for collaborations and shared efforts. But even together, these datasets do not provide enough training data for models much larger than a few billion parameters. So in order to expand the options for open model training, we still need more open data...

Based on an analysis of 1 million user interactions with ChatGPT, the plurality of user requests are for creative compositions... The kind of content we actually need — like creative writing — is usually tied up in copyright restrictions. Common Corpus tackles these challenges through five carefully curated collections...
Last week AMD also released its first series of fully open 1 billion parameter language models, AMD OLMo.

And last month VentureBeat reported that the non-profit Allen Institute for AI had unveiled Molmo, "an open-source family of state-of-the-art multimodal AI models which outpeform top proprietary rivals including OpenAI's GPT-4o, Anthropic's Claude 3.5 Sonnet, and Google's Gemini 1.5 on several third-party benchmarks."

jojowombl shares a report from The Guardian: Legal documents released in ongoing US litigation between NSO Group and WhatsApp have revealed for the first time that the Israeli cyberweapons maker -- and not its government customers -- is the party that "installs and extracts" information from mobile phones targeted by the company's hacking software. The new details were contained in sworn depositions from NSO Group employees, portions of which were published for the first time on Thursday.

It comes five years after WhatsApp, the popular messaging app owned by Facebook, first announced it was filing suit against NSO. The company, which was blacklisted by the Biden administration in 2021, makes what is widely considered the world's most sophisticated hacking software, which -- according to researchers -- has been used in the past in Saudi Arabia, Dubai, India, Mexico, Morocco and Rwanda. [...] At the heart of the legal fight was an allegation by WhatsApp that NSO had long denied: that it was the Israeli company itself, and not its government clients around the world, who were operating the spyware. NSO has always said that its product is meant to be used to prevent serious crime and terrorism, and that clients are obligated not to abuse the spyware. It has also insisted that it does not know who its clients are targeting. [...]

To make its case, WhatsApp was allowed by Judge Phyllis Hamilton to make its case, including citing depositions that have previously been redacted and out of public view. In one, an NSO employee said customers only needed to enter a phone number of the person whose information was being sought. Then, the employee said, "the rest is done automatically by the system." In other words, the process was not operated by customers. Rather NSO alone decided to access WhatsApp's servers when it designed (and continuously upgraded) Pegasus to target individuals' phones. A spokesperson for NSO, Gil Lainer, said in a statement: "NSO stands behind its previous statements in which we repeatedly detailed that the system is operated solely by our clients and that neither NSO nor its employees have access to the intelligence gathered by the system. We are confident that these claims, like many others in the past, will be proven wrong in court, and we look forward to the opportunity to do so."

Economist: Britain has long been a pioneer in telecoms. In 1837 it built the world's first commercial telegraph; the first transatlantic call was placed from London in 1927; in 1992 a British programmer sent the first text message to a mobile phone. Today it lags rather than leads. According to figures provided to The Economist by Opensignal, a research firm, Britain ranks 46th for download speeds out of the 56 developed and developing countries for which there are data. That gives it the worst mobile service in the rich world.

Some of this is due to demand. Over the past three years data usage on mobile devices has doubled as people stream films and play games. The busiest parts of cities often lack mobile reception because the system is at capacity. But mainly it is an issue of supply. British users of 5G are only on it 11% of the time. That puts Britain 43rd out of the 56 countries. This lacklustre performance is caused by a combination of government U-turns, insufficient investment and sclerotic planning.

Brazilian officials are scrambling to control a gambling boom that has led some citizens to take out loans with interest rates as high as 438% to fund their betting habits, sparking concerns about household debt levels.

The surge in online betting has doubled Brazil's gambling population to 52 million in six months, with the central bank estimating monthly gambling spending between 18-21 billion reais ($3.1-3.6 billion) through August 2024. Central Bank President Roberto Campos Neto said lower-income families are disproportionately affected, with 20% of government social program payments in August directed to online gambling sites.

The Finance Ministry has accelerated regulatory measures, requiring over 100 betting companies to submit operating paperwork ahead of schedule. New rules starting January 1 will allow authorities to limit bet amounts, block payment systems, and monitor for money laundering. President Luiz Inacio Lula da Silva recently raised concerns at the UN about gambling's impact on Brazil's poorest citizens, while officials are moving to ban credit card use for betting and restrict gambling advertisements.

An anonymous reader quotes a report from Bloomberg: The Australian government plans to enact laws requiring big tech firms to protect its citizens online, the latest move by the center-left Labor administration to crack down on social media including through age limits and curbs on misinformation. Communications Minister Michelle Rowland announced the government's plan for a legislated Digital Duty of Care in Australia on Wednesday night, saying it aligned with similar laws in the UK and European Union. "It is now time for industry to show leadership, and for social media to recognize it has a social responsibility," Rowland said in a speech in Sydney announcing the measures. It would "keep users safe and help prevent online harms."

In response to the laws, Facebook and Instagram operator Meta Platforms Inc. called for the restrictions to be handled by app stores, such as those run by Google and Apple Inc., rather than the platforms themselves. The government has ignored those requests, but has yet to announce what fines companies would face or what age verification information will need to be provided. At the same time, Albanese has moved forward controversial laws to target misinformation and disinformation online, which opponents have labeled an attack on freedom of speech. Earlier this month, Albanese said the government would legislate for a ban on social media for children under 16, a policy the government says is world-leading. "Social media is doing harm to our kids and I'm calling time on it," Albanese told a news conference.

Datacenter operators are increasingly turning to small modular reactors (SMRs) like those developed by Oklo to meet growing energy demands. According to The Register, Oklo has secured commitments from two major datacenter providers for 750 MW of power, pending regulatory approvals. It brings the firm's planned nuclear build-out to 2.1 gigawatts. From the report: Oklo's designs are, from what we understand, inspired by the Experimental Breeder Reactor II (EBR-II) and utilize liquid-metal cooling. They are capable of producing between 15MW and 50MW of power, depending on the configuration. That means Oklo's datacenter customers plan to deploy somewhere between 15 and 50 of the reactors to satisfy their thirst for electricity. However, they may be waiting a while.

According to Oklo's website, the nuclear startup hopes to bring its first plant online before the end of the decade. Before that can happen, though, Oklo will need to obtain approval from the Nuclear Regulatory Commission -- something for which it says it's already submitted applications. In 2022, the watchdog rejected an Oklo plan to build a small atomic reactor in Idaho, citing "significant information gaps" on safety-related measures.

That said, Oklo has lately received support from US government agencies including the Department of Energy (DoE), which has awarded a site use permit, while Idaho National Laboratory -- home of EBR-II -- has provided fuel material to support the efforts. Speaking of fuel, Oklo's designs may not suffer from the challenges other SMR startups, like Terrapower, have encountered. Oklo's designs are intended to run on recycled nuclear waste products from traditional reactors. In fact, the startup is currently working with DoE national labs to develop new fuel recycling technologies. Oklo hopes to bring a commercial-scale recycling plan online by the early 2030s.

quonset shares a report from CNN: Between August 2022 and January 2024, hundreds of swatting calls were made across the country targeting religious institutions, government offices, schools, and random people. Authorities were finally able to track down the criminal, Alan Fillon, who entered the plea to four counts of making interstate threats to injure the person of another, the US Attorney's Office for the Middle District of Florida said in a news release. He faces up to five years in prison on each count. A sentencing date has not yet been set.

The US Attorney's Office said Filion made more than 375 swatting and threat calls from August 2022 to January 2024. Those calls included ones in which he claimed to have planted bombs in targeted locations or threatened to detonate bombs and/or conduct mass shootings at those locations, prosecutors said. He targeted religious institutions, high schools, colleges and universities, government officials and people across the United States. Filion was 16 at the time he placed the majority of the calls.

An anonymous reader quotes a report from Data Center Dynamics: The Japanese government is planning to invest approximately $65 billion to support the country's semiconductor and AI industries. The initiative, which will run until the end of the decade, is expected to generate ~$104 billion in public and private investment during the period. According to a report from Reuters, this new round of funding will specifically target state-backed chip foundry Rapidus and other AI chip suppliers.

Rapidus was founded in November 2022 when the Japanese government and eight Japanese technology and automotive firms, including SoftBank, Sony, and NTT, invested more than $500 million to launch the business. Speaking at a news conference this week, Japanese Prime Minister Shigeru Ishiba did not provide any information about how the venture would be financed but said the government would not issue deficit-covering bonds. Japan's government also said it won't raise taxes to finance the $65 billion plan.

Washington Post: The Consumer Financial Protection Bureau has taken steps to place Google under formal federal supervision, an extraordinary move that could subject the technology giant to the regular inspections and other rigorous monitoring that the government imposes on major banks.

Google has fiercely resisted the idea over months of highly secretive talks, according to two people familiar with the discussions, who spoke on the condition of anonymity to describe them -- setting up what may ultimately be a major legal clash with vast implications for the CFPB's powers in the digital age.

The exact scope of the CFPB's concerns is not clear, and its order does not appear to be final. The political fate of the bureau's work under Director Rohit Chopra is also in doubt, as the watchdog agency braces for potentially significant changes to its leadership and agenda with the return of President-elect Donald Trump to the White House.

Formed in the aftermath of the 2008 financial crisis, the CFPB has broad powers to protect consumers from unfair, deceptive or predatory financial practices. That includes the ability to place certain firms under supervision, a status that can afford regulators direct access to the company's internal records to ensure their activities are sound -- and seek fixes if they are not.

During a public joint hearing today titled "Unidentified Anomalous Phenomena: Exposing the Truth," four experts testified that the U.S. is running secret UAP programs, including crash retrieval and reverse-engineering programs for advanced nonhuman technology. Although the Pentagon maintains there's no evidence of alien spacecraft, witnesses like Luis Elizondo and Michael Gold argue that UAPs represent an intelligence enigma and call for open, stigma-free study to address potential security concerns and unknown scientific possibilities. NPR reports: Tim Gallaudet, retired rear admiral, U.S. Navy; CEO of Ocean STL Consulting, LLC
"Confirmation that UAPs are interacting with humanity came for me in January 2015," Gallaudet said in his written testimony (PDF). He describes being part of a pre-deployment naval exercise off the U.S. East Coast that culminated in the famous "Go Fast" video, in which a Navy F/A-18 jet's sensors recorded "an unidentified object exhibiting flight and structural characteristics unlike anything in our arsenal." He was among a group of commanders involved in the exercise who received an email containing the video, which was sent by the operations officer of Fleet Forces Command, Gallaudet said. "The very next day, the email disappeared from my account and those of the other recipients without explanation," he said.

Luis Elizondo, author and former Department of Defense official
Elizondo's written testimony (PDF) was brief and alleged that a secretive arms race is playing out on the global stage. "Let me be clear: UAP are real," he wrote. "Advanced technologies not made by our Government -- or any other government -- are monitoring sensitive military installations around the globe. Furthermore, the U.S. is in possession of UAP technologies, as are some of our adversaries." Elizondo is a former intelligence officer who later "managed a highly sensitive Special Access Program on behalf of the White House and the National Security Council," according to his official bio (PDF). "By 2012, [Elizondo] was the senior ranking person of the DOD's Advanced Aerospace Threat Identification Program, a secretive Pentagon unit that studied unidentified anomalous phenomena," his bio states, adding that he resigned in 2017.

Michael Gold, former NASA associate administrator of space policy and partnerships; member of NASA UAP Independent Study Team
Gold's written testimony (PDF) stressed the need for government agencies and academics to "overcome the pernicious stigma that continues to impede scientific dialogue and open discussions" about unexplained phenomena. "As the saying goes, the truth is out there," Gold said, "we just need to be bold enough and brave enough to face it."

Michael Shellenberger, founder of Public, a news outlet on the Substack platform
Shellenberger's testimony (PDF) ran to some 214 pages, including a lengthy timeline of UAP reports from 1947 to 2023. Shellenberger pressed the White House and Congress to act, calling for the adoption of UAP transparency legislation and cutting funds for any related programs that aren't disclosed to lawmakers. "UAP transparency is bi-partisan and critical to our national security," his written testimony stated. You can watch the proceeding here.

An anonymous reader shares a report: Officials inside the Secret Service clashed over whether they needed a warrant to use location data harvested from ordinary apps installed on smartphones, with some arguing that citizens have agreed to be tracked with such data by accepting app terms of service, despite those apps often not saying their data may end up with the authorities, according to hundreds of pages of internal Secret Service emails obtained by 404 Media.

The emails provide deeper insight into the agency's use of Locate X, a powerful surveillance capability that allows law enforcement officials to follow a phone, and person's, precise movements over time at the click of a mouse. In 2023, a government oversight body found that the Secret Service, Customs and Border Protection, and Immigration and Customs Enforcement all used their access to such location data illegally. The Secret Service told 404 Media in an email last week it is no longer using the tool. "If USSS [U.S. Secret Service] is using Locate X, that is most concerning to us," one of the internal emails said. 404 Media obtained them and other documents through a Freedom of Information Act (FOIA) request with the Secret Service.

An anonymous reader quotes a report from The Register: Royal assent was granted to two right to repair bills last week that amend Canada's Copyright Act to allow the circumvention of technological protection measures (TPMs) if this is done for the purposes of "maintaining or repairing a product, including any related diagnosing," and "to make the program or a device in which it is embedded interoperable with any other computer program, device or component." The pair of bills allow device owners to not only repair their own stuff regardless of how a program is written to prevent such non-OEM measures, but said owners can also make their devices work with third-party components without needing to go through the manufacturer to do so.

Bills C-244 (repairability) and C-294 (interoperability) go a long way toward advancing the right to repair in Canada and, as iFixit pointed out, are the first federal laws anywhere that address how TPMs restrict the right to repair -- but they're hardly final. TPMs can take a number of forms, from simple administrative passwords to encryption, registration keys, or even the need for a physical object like a USB dongle to unlock access to copyrighted components of a device's software. Most commercially manufactured devices with proprietary embedded software include some form of TPM, and neither C-244 nor C-294 place any restrictions on the use of such measures by manufacturers. As iFixit points out, neither Copyright Act amendments do anything to expand access to the tools needed to circumvent TPMs. That puts Canadians in a similar position to US repair advocates, who in 2021 saw the US Copyright Office loosen DMCA restrictions to allow limited repairs of some devices despite TPMs, but without allowing access to the tools needed to do so. [...]

Canadian Repair Coalition co-founder Anthony Rosborough said last week that the new repairability and interoperability rules represent considerable progress, but like similar changes in the US, don't actually amount to much without the right to distribute tools. "New regulations are needed that require manufacturers and vendors to ensure that products and devices are designed with accessibility of repairs in mind," Rosborough wrote in an op-ed last week. "Businesses need to be able to carry out their work without the fear of infringing various intellectual property rights."

Longtime Slashdot reader thephydes writes: The hearing will go ahead on November 13 at 11:30 ET (16:30 GMT). Apparently, it will "further pull back the curtain on secret UAP research programs conducted by the U.S. government, and undisclosed findings they have yielded," according to a House statement. It's driven by two republicans, Nancy Mace (R-S.C.) and Glenn Grothman (R-Wis.), who say: "Americans deserve to understand what the government has learned about UAP sightings, and the nature of any potential threats these phenomena pose. We can only ensure that understanding by providing consistent, systemic transparency. We look forward to hearing from expert witnesses on ways to shed more light and bring greater accountability to this issue." "Expert witnesses in the hearing will include Luis Elizondo, a decorated former counterintelligence officer who has claimed for years that the U.S. government is hiding knowledge of UAP, including materials recovered from crashed flying saucers," reports Space.com. "The House hearing will also include Tim Gallaudet, a retired U.S. Navy Rear Admiral who observed unidentified submersible objects, arguing that 'these underwater anomalies jeopardize US maritime security.'"

"Other speakers at the hearing include journalist Michael Shellenberger, who has also claimed the U.S. government is hiding UFO crash retrieval programs, and former NASA Associate Administrator of Space Policy and Partnerships Michael Gold, who is a member of NASA's independent UAP study team."

Pakistan's Punjab province banned most outdoor activities and ordered shops, markets and malls in some areas to close early from Monday to curb illnesses caused by intense air pollution. From a report: The province has closed educational institutions and public spaces like parks and zoos until Nov. 17 in places including Lahore, the world's most polluted city in terms of air quality, according to Swiss group IQAir's live ratings.

The districts of Lahore, Multan, Faisalabad and Gujranwala have seen an unprecedented rise in patients with respiratory diseases, eye and throat irritation, and pink eye disease, the Punjab government said in an order issued late on Sunday. The new restrictions will also remain in force until Nov. 17. "The spread of conjunctivitis/ pink eye disease due to bacterial or viral infection, smoke, dust or chemical exposure is posing a serious and imminent threat to public health," the Punjab government said.

While outdoor activities including sports events, exhibitions and festivals, and dining at restaurants have been prohibited, "unavoidable religious rites" are exempt from this direction, the order said. Outlets like pharmacies, oil depots, dairy shops and fruit and vegetable shops have similarly been exempted from the directions to close by 8 p.m. local time. Lahore's air quality remained hazardous on Monday, with an index score of more than 600, according to IQAir, but this was significantly lower than the 1,900 that it touched in places earlier this month. A score of 0-50 is considered good.

The Biden administration plans to support a controversial cybercrime treaty at the United Nations this week despite concerns that it could be misused by authoritarian regimes, Bloomberg News reported Monday, citing senior government officials. From the report: The agreement would be the first legally binding UN agreement on cybersecurity and could become a global legal framework for countries to cooperate on preventing and investigating cybercriminals. However, critics fear it could be used by authoritarian states to try to pursue dissidents overseas or collect data from political opponents. Still, the officials said there are persuasive reasons to support the treaty. For instance, it would advance the criminalization of child sexual-abuse material and nonconsensual spreading of intimate images, they said.

In addition, the wider involvement of member states would make cybercrime and electronic evidence more available to the US, one official said. If all the members sign the agreement, it would update extradition treaties and provide more opportunities to apprehend cybercriminals and have them extradited, the official added. Hundreds of submissions from advocacy groups and other parties criticized US involvement in the agreement. The US plans to strictly enforce human rights and other safeguards in the treaty, the officials said, adding that the Department of Justice would closely scrutinize requests and refuse to provide any assistance that was inconsistent with the agreement.

On Tuesday Massachusetts voted to become the first state to allow gig-working drivers to join labor unions, reports WBUR: Since these gig workers are classified as independent contractors, federal law allowing employees the right to unionize does not apply to them. With the passage of this ballot initiative, Massachusetts is the first state to give ride-hailing drivers the ability to collectively bargain over working conditions.
Supporters have said the ballot measure "could provide a model for other states to let Uber and Lyft drivers unionize," reports Reuters, "and inspire efforts to organize them around the United States." Roxana Rivera, assistant to the president of 32BJ SEIU, an affiliate of the Service Employees International Union, that had spearheaded a campaign to pass the proposal, said its approval shows that Massachusetts voters want drivers to have a meaningful check against the growing power of app-based companies... The Massachusetts vote was the latest front in a years-long battle in the United States over whether ride-share drivers should be considered to be independent contractors or employees entitled to benefits and wage protections. Studies have shown that using contractors can cost companies as much as 30% less than employees.

Drivers for Uber and Lyft, including approximately 70,000 in Massachusetts, do not have the right to organize under the National Labor Relations Act... Under the Massachusetts measure, drivers can form a union after collecting signatures from at least 25% of active drivers in Massachusetts, and companies can form associations to allow them to jointly negotiate with the union during state-supervised talks.
But the Boston Globe points out that the measure " divided labor advocates in Massachusetts, some of whom worry it would in fact be a step backward in the lengthy fight to boost the rights of gig workers." Those concerns led the state's largest labor organization, the AFL-CIO, to remain neutral. But two unions backing the effort, the SEIU 32BJ and the International Association of Machinists, say allowing drivers to unionize, even if not as full employees, will help provide urgently needed worker protections and better pay and safety standards.

Tuesday at California's Vandenberg Space Force base, the U.S. launched a Minuteman III missile, "in an important test of the weapon's ability to strike its targets with multiple warheads," according to Air and Space Forces magazine: The Minuteman III missiles that form a critical leg of the U.S. nuclear triad each carry one nuclear-armed reentry vehicle. But the missile that was tested carried three test warheads... The intercontinental ballastic missile (ICBM) test was controlled by an airborne command post in a test of the U.S. ability to launch its nuclear deterrent from a survivable platform.... Gen. Thomas A. Bussiere, the commander of Air Force Global Strike Command, said in a release: "An airborne launch validates the survivability of our ICBMs, which serve as the strategic backstop of our nation's defense and defense of allies and partners...."

The three test reentry vehicles — one high-fidelity Joint Test Assembly, which carries non-nuclear explosives, and two telemetry Joint Test Assembly objects — struck the Reagan Test Site near the Kwajalein Atoll in the Marshall Islands roughly 30 minutes later after launch, a flight of about 4,200 miles. "They make up essentially a mock warhead," Col. Dustin Harmon, the commander of the 377th Test and Evaluation Group, the nation's operational ICBM test unit, said in an interview with Air & Space Forces Magazine. "There's two different types. One is telemetered, so it's got a radio transmitter in it, it's got antennas, gyroscopes, accelerometers — all the things that can sense motion and movement. And we fly those or we can put one in there that's called a high-fidelity. That is assembled much like an actual weapon would be, except we use surrogate materials, and so we want it to fly similarly to an actual weapon. ... It has the explosives in it that a normal warhead would to drive a detonation, but there's nothing to drive...."

The U.S. government formally notified Russia in advance of the launch in accordance with a 1988 bilateral agreement. More than 145 countries were also provided with advance notice of the launch under the Hague Code of Conduct — an international understanding on launch notifications. The U.S. also provided advance notice to China, a DOD spokesperson told Air & Space Forces Magazine. China notified the U.S. of an ICBM launch over the Pacific Ocean in September. There is no formal agreement between Washington and Beijing that requires such notifications, but each side provided them to avoid miscalculations.
Test launches happen three times a year, according to the article, yielding "several gigabytes of data" about reentry vehicles, subsystems, and payloads. "There are 400 Minuteman III missiles currently in service across Colorado, Montana, Nebraska, North Dakota, and Wyoming."

Thanks to long-time Slashdot reader SonicSpike for sharing the article.

An anonymous reader quotes a report from TechCrunch: The FBI is warning that hackers are obtaining private user information — including emails and phone numbers — from U.S.-based tech companies by compromising government and police email addresses to submit "emergency" data requests. The FBI's public notice filed this week is a rare admission from the federal government about the threat from fraudulent emergency data requests, a legal process designed to help police and federal authorities obtain information from companies to respond to immediate threats affecting someone's life or property. The abuse of emergency data requests is not new, and has been widely reported in recent years. Now, the FBI warns that it saw an "uptick" around August in criminal posts online advertising access to or conducting fraudulent emergency data requests, and that it was going public for awareness.

"Cyber-criminals are likely gaining access to compromised US and foreign government email addresses and using them to conduct fraudulent emergency data requests to US based companies, exposing the personal information of customers to further use for criminal purposes," reads the FBI's advisory. [...] The FBI said in its advisory that it had seen several public posts made by known cybercriminals over 2023 and 2024, claiming access to email addresses used by U.S. law enforcement and some foreign governments. The FBI says this access was ultimately used to send fraudulent subpoenas and other legal demands to U.S. companies seeking private user data stored on their systems. The advisory said that the cybercriminals were successful in masquerading as law enforcement by using compromised police accounts to send emails to companies requesting user data. In some cases, the requests cited false threats, like claims of human trafficking and, in one case, that an individual would "suffer greatly or die" unless the company in question returns the requested information.

The FBI said the compromised access to law enforcement accounts allowed the hackers to generate legitimate-looking subpoenas that resulted in companies turning over usernames, emails, phone numbers, and other private information about their users. But not all fraudulent attempts to file emergency data requests were successful, the FBI said. The FBI said in its advisory that law enforcement organizations should take steps to improve their cybersecurity posture to prevent intrusions, including stronger passwords and multi-factor authentication. The FBI said that private companies "should apply critical thinking to any emergency data requests received," given that cybercriminals "understand the need for exigency."

An anonymous reader quotes a report from Ars Technica: Anthropic has announced a partnership with Palantir and Amazon Web Services to bring its Claude AI models to unspecified US intelligence and defense agencies. Claude, a family of AI language models similar to those that power ChatGPT, will work within Palantir's platform using AWS hosting to process and analyze data. But some critics have called out the deal as contradictory to Anthropic's widely-publicized "AI safety" aims. On X, former Google co-head of AI ethics Timnit Gebru wrote of Anthropic's new deal with Palantir, "Look at how they care so much about 'existential risks to humanity.'"

The partnership makes Claude available within Palantir's Impact Level 6 environment (IL6), a defense-accredited system that handles data critical to national security up to the "secret" classification level. This move follows a broader trend of AI companies seeking defense contracts, with Meta offering its Llama models to defense partners and OpenAI pursuing closer ties with the Defense Department. In a press release, the companies outlined three main tasks for Claude in defense and intelligence settings: performing operations on large volumes of complex data at high speeds, identifying patterns and trends within that data, and streamlining document review and preparation.

While the partnership announcement suggests broad potential for AI-powered intelligence analysis, it states that human officials will retain their decision-making authority in these operations. As a reference point for the technology's capabilities, Palantir reported that one (unnamed) American insurance company used 78 AI agents powered by their platform and Claude to reduce an underwriting process from two weeks to three hours. The new collaboration builds on Anthropic's earlier integration of Claude into AWS GovCloud, a service built for government cloud computing. Anthropic, which recently began operations in Europe, has been seeking funding at a valuation up to $40 billion. The company has raised $7.6 billion, with Amazon as its primary investor.

The FBI is warning that hackers are obtaining private user information -- including emails and phone numbers -- from U.S.-based tech companies by compromising government and police email addresses to submit "emergency" data requests. From a report: The FBI's public notice filed this week is a rare admission from the federal government about the threat from fraudulent emergency data requests, a legal process designed to help police and federal authorities obtain information from companies to respond to immediate threats affecting someone's life or property.

The abuse of emergency data requests is not new, and has been widely reported in recent years. Now, the FBI warns that it saw an "uptick" around August in criminal posts online advertising access to or conducting fraudulent emergency data requests, and that it was going public for awareness. "Cyber-criminals are likely gaining access to compromised U.S. and foreign government email addresses and using them to conduct fraudulent emergency data requests to U.S. based companies, exposing the personal information of customers to further use for criminal purposes," reads the FBI's advisory.