Latest Tech News

Stay updated with the latest in technology, AI, cybersecurity, and more

Filtered by: researchers Clear Filter

Researchers from OpenAI, Anthropic, Meta, and Google issue joint AI safety warning - here's why

Andriy Onufriyenko / Getty Images Over the last year, chain of thought (CoT) -- an AI model's ability to articulate its approach to a query in natural language -- has become an impressive development in generative AI, especially in agentic systems. Now, several researchers agree it may also be critical to AI safety efforts. On Tuesday, researchers from competing companies including OpenAI, Anthropic, Meta, and Google DeepMind, as well as institutions like the Center for AI Safety, Apollo Resea

OpenAI, Google, and Meta Researchers Warn We May Lose the Ability to Track AI Misbehavior

Over 40 scientists from the world’s leading AI institutions, including OpenAI, Google DeepMind, Anthropic, and Meta, have come together to call for more research in a particular type of safety monitoring that allows humans to analyze how AI models “think.” The scientists published a research paper on Tuesday that highlighted what is known as chain of thought (CoT) monitoring as a new yet fragile opportunity to boost AI safety. The paper was endorsed by prominent AI figures like OpenAI co-founde

Top AI Researchers Concerned They’re Losing the Ability to Understand What They’ve Created

Researchers from OpenAI, Google DeepMind, and Meta have joined forces to warn about what they're building. In a new position paper, 40 researchers spread across those four companies called for more investigation of AI powered by so-called "chains-of-thought" (CoT), the "thinking out loud" process that advanced "reasoning" models — the current vanguard of consumer-facing AI — use when they're working through a query. As those researchers acknowledge, CoTs add a certain transparency into the inn

OpenAI, Google DeepMind and Anthropic sound alarm: ‘We may be losing the ability to understand AI’

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Scientists from OpenAI, Google DeepMind, Anthropic and Meta have abandoned their fierce corporate rivalry to issue a joint warning about artificial intelligence safety. More than 40 researchers across these competing companies published a research paper today arguing that a brief window to monitor AI reasoning could close forever — and soon

North Korean XORIndex malware hidden in 67 malicious npm packages

North Korean threat actors planted 67 malicious packages in the Node Package Manager (npm) online repository to deliver a new malware loader called XORIndex to developer systems. The packages collectively count more than 17,000 downloads and were discovered by researchers at package security platform Socket, who assess them to be part of the continued Contagious Interview operation. Socket researchers say that the campaign follows threat activity detected since April. Last month, the same acto

Nvidia chips become the first GPUs to fall to Rowhammer bit-flip attacks

Nvidia is recommending a mitigation for customers of one of its GPU product lines that will degrade performance by up to 10 percent in a bid to protect users from exploits that could let hackers sabotage work projects and possibly cause other compromises. The move comes in response to an attack a team of academic researchers demonstrated against Nvidia’s RTX A6000, a widely used GPU for high-performance computing that’s available from many cloud services. A vulnerability the researchers discove

This Common Pain Med Could Be Raising Dementia Risk

Chronic pain can be a debilitating experience. A common medication used to manage the condition may come with its own serious dangers, however. Recent research finds an upsetting link between the drug gabapentin and a higher risk of dementia. Scientists at Case Western Reserve University led the study, which examined the medical records of people with chronic low back pain. People taking gabapentin for their back were significantly more likely to develop dementia over time compared to non-users

Bold Mission to Hunt for Aliens on Venus Is Happening

A UK-based mission is aiming to settle, once and for all, whether life exists on Venus. The mission plans to send a probe to the planet in search of microbial life, not on the surface, of course, but in the Venusian clouds. Over the past half-decade, scientists have detected the presence of phosphine and ammonia—two potential signs of biological activity—in Venus’s clouds. On Earth, both gases are produced only by biological activity and industrial processes, and scientists are unsure of their

Scientists Gene Hack Mice So Their Livers Produce Their Own Ozempic-Like Drug

Image by Remi Benali/Getty Images Rx/Medicines Scientists have gene-hacked mice to produce their own Ozempic-like drugs — possibly, and provocatively, perhaps paving a path for humans to do so themselves one day. In a new study published in the journal Communications Medicine, researchers from Japan's University of Osaka successfully gene-edited mice livers to produce exenatide, a first-generation diabetes drug and predecessor to now-trendy jabs like Ozempic and Mounjaro. Using lab mice that

'Starter packs' have played a central role in Bluesky's rapid growth

“Vital onboarding strategy for the emerging social media systems” “Our findings go beyond Bluesky: they point to a new framework for launching successful social platforms,” said Dr Onur Ascigil, Lecturer in Computer Science at Lancaster University and Principal Investigator of the study. “Starter packs are becoming a vital onboarding strategy for the emerging social media systems that are seeking to attract users from dominant platforms.” The researchers believe their findings could help platfo

Chinese Scientists Create Cyborg Bees That Can Be Controlled Like Drones for Undercover Military Missions

Chinese Scientists Create Cyborg Bees That Can Be Controlled Like Drones for Undercover Military Missions A tiny backpack relays commands straight to their brain. Seal Team Bee Researchers at the Beijing Institute of Technology have turned innocent bees into cyborgs that can be controlled via a 74-milligram insect brain controller. As the South China Morning Post reports, the controller pierces the bee's tiny brain with three needles and uses signals sent via electronic pulses to make it fly

AI chatbot’s simple ‘123456’ password risked exposing personal data of millions of McDonald’s job applicants

In Brief Security researchers found that they could access the personal information of 64 million people who had applied for a job at McDonald’s, in large part by logging into the company’s AI job hiring chatbot with the username and password “123456.” Ian Carroll and Sam Curry wrote in a blog post that “during a cursory security review of a few hours,” they found the password issue and another simple security vulnerability in an internal API, which allowed access to job applicants’ past conve

A Bold Mission to Hunt for Aliens on Venus Is Actually Happening

A UK-based mission is aiming to settle, once and for all, whether life exists on Venus. The mission plans to send a probe to the planet in search of microbial life, not on the surface, of course, but in the Venusian clouds. Over the past half-decade, scientists have detected the presence of phosphine and ammonia—two potential signs of biological activity—in Venus’s clouds. On Earth, both gases are produced only by biological activity and industrial processes, and scientists are unsure of their

“Things we’ll never know” science fair highlights US’s canceled research

Washington, DC—From a distance, the gathering looked like a standard poster session at an academic conference, with researchers standing next to large displays of the work they were doing. Except in this case, it was taking place in the Rayburn House Office Building on Capitol Hill, and the researchers were describing work that they weren’t doing. Called "The things we’ll never know," the event was meant to highlight the work of researchers whose grants had been canceled by the Trump administrat

“Things we’ll never know” science fair highlights US’ canceled research

Washington, DC — From a distance, the gathering looked like a standard poster session at an academic conference, with researchers standing next to large displays of the work they were doing. Except in this case, it was taking place in the Rayburn House Office Building on Capitol Hill, and the researchers were describing work that they weren’t doing. Called "The things we’ll never know," the event was meant to highlight the work of researchers whose grants had been canceled by the Trump administr

Your Fitbit Could Become Your Post-Surgery Best Friend

A Fitbit a day just might help keep your post-surgery woes at bay. Research today finds that wearable data can predict children’s risk of health problems following a removed appendix. Scientists in Chicago conducted the study, which equipped over a hundred children with Fitbits after their appendectomy. Using a specially designed algorithm, the Fitbits accurately detected whether children would develop postoperative complications, often days before they were formally diagnosed. The findings sug

Chinese researchers unveil MemOS, the first ‘memory operating system’ that gives AI human-like recall

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now A team of researchers from leading institutions including Shanghai Jiao Tong University and Zhejiang University has developed what they’re calling the first “memory operating system” for artificial intelligence, addressing a fundamental limitation that has hindered AI systems from achieving human-like persistent memory and learning. The sy

Scientists Find Alarming Link Between AI Use and Psychopathy

Artificial intelligence use has been associated with everything from fear of judgment and loneliness to misogyny and illiteracy — a baffling array of outcomes that's often alarming, but defies easy categorization. Now the plot thickens. In a new study published in the journal BMC Psychology, South Korean scientists surveyed 504 college-level Chinese art students and found that the ones who exhibited higher rates of narcissism, psychopathy, and Machiavellianism were more likely rely on ChatGPT a

'Batavia' Windows spyware campaign targets dozens of Russian orgs

A previously undocumented spyware called ‘Batavia’ has been targeting large industrial enterprises in Russia in a phishing email campaign that uses contract-related lures. The researchers believe the operation has been active since at least last year in July and is ongoing. Based on telemetry data, the phishing emails delivering Batavia have reached employees at several dozen Russian organizations have been targeted. Since January 2025, the campaign has increased in intensity and peaked toward

Lost for 300 Years, Pirate-Plundered Treasure Ship Discovered off Madagascar Coast

In 1721, pirates attacked and seized a Portuguese ship carrying a massive trove of treasure en route to Lisbon. Now, researchers believe they’ve discovered its remains off the coast of Madagascar. The discovery comes from two researchers from the Center for Historic Shipwreck Preservation in Massachusetts, who have conducted several studies on the wreckage over the last 16 years. They say new clues have revealed the ship’s identity as the Nossa Senhora do Cabo, a 700-ton warship. Their findings

Scientists Finally Sequenced the First Ancient Egyptian Genome

Scientists have, for the first time, sequenced the entire genome of an ancient Egyptian who lived approximately 4,500 to 4,800 years ago. The feat was achieved by a team of researchers at the Francis Crick Institute and Liverpool John Moores University, who published their findings in Nature. According to the study, the ancient individual’s genetic ancestry traces back to populations in both North Africa and West Asia, shedding light on the genetic diversity of early Egyptians. Researchers fir

Scientists Uncover Exercise Lifehack: Go to Bed

As if you needed another reason to envy—or loathe—morning people. Research this week shows that people who go to bed early are more likely to be physically active than those who crave the night. Scientists at Monash University in Australia led the study, which objectively examined people’s sleeping and exercise habits. Compared to late-night and typical sleepers, people who went to bed early tended to perform more physical activity the following day, they found. The findings also suggest that t

A Pro-Russia Disinformation Campaign Is Using Free AI Tools to Fuel a ‘Content Explosion’

A pro-Russia disinformation campaign is leveraging consumer artificial intelligence tools to fuel a “content explosion” focused on exacerbating existing tensions around global elections, Ukraine, and immigration, among other controversial issues, according to new research published last week. The campaign, known by many names including Operation Overload and Matryoshka (other researchers have also tied it to Storm-1679), has been operating since 2023 and has been aligned with the Russian govern

Millions of Brother Printers Are Full of Hackable Bugs

Brother makes some solid, reliable printers. Indeed, for several years running, The Verge named it the best printer you should buy. Unfortunately, the company’s devices appear to be riddled with new zero-day bugs that could allow a savvy cybercriminal to hijack them. The vulnerabilities were discovered by cybersecurity firm Rapid7, which published a blog about the bugs last week. The blog explains that, after some research, Rapid7’s cyber pros came across a total of eight new zero-day vulnerabi

OpenAI reportedly ‘recalibrating’ compensation in response to Meta hires

In Brief With Meta successfully poaching a number of its senior researchers, an OpenAI executive reportedly reassured team members Saturday that company leadership has not “been standing idly by.” “I feel a visceral feeling right now, as if someone has broken into our home and stolen something,” Chief Research Officer Mark Chen wrote in a Slack memo obtained by Wired. In response to what appears to be a Meta hiring spree, Chen said that he, CEO Sam Altman, and other OpenAI leaders have been w

Scientists Intrigued to Discover That Human Brains Are Glowing Faintly

Image by Getty / Futurim Developments Scientists have some exciting news: your brain is likely glowing, whether you can see it or not. The news comes from researchers at Algoma University in Ontario, who found evidence that the human brain, of all things, possesses luminescent properties. Essentially, they found that as the brain metabolizes energy, it releases super-faint traces of visible light. Called ultra-weak photon emissions (UPEs), the flashes of light are emitted when electrons break

Bluetooth flaws could let hackers spy through your microphone

Vulnerabilities affecting a Bluetooth chipset present in more than two dozen audio devices from ten vendors can be exploited for eavesdropping or stealing sensitive information. Researchers confirmed that 29 devices from Beyerdynamic, Bose, Sony, Marshall, Jabra, JBL, Jlab, EarisMax, MoerLabs, and Teufel are affected. The list of impacted products includes speakers, earbuds, headphones, and wireless microphones. The security problems could be leveraged to take over a vulnerable product and on

Notorious Fungus Blamed for ‘Mummy’s Curse’ Is Now a Promising Cancer Treatment

In the 1920s, a number of workers on the excavation team that uncovered King Tutankhamun’s tomb met untimely deaths. Five decades later, 10 out of 12 scientists died after entering the tomb of the 15th-century Polish King Casimir IV. In both cases, researchers suggested that fungal spores could have played a role in the mysterious deaths, specifically identifying the fungus Aspergillus flavus within the Polish burial. A. flavus is now making a comeback, but not as a reawakened killer from ancie

Scientists Launch Wild New Project to Build a Human Genome From Scratch

A team of UK-based researchers is going where no scientist has dared to go—writing artificial human DNA from scratch. They’re hoping the project will answer fundamental questions about the human genome and transform our understanding of health and disease. But the research topic is, for obvious reasons, controversial. Scientists have largely steered clear of trying to create full synthetic human genomes, wary of propelling us into a dystopian, Gattaca-esque future full of designer babies. Now,

This Prehistoric Trick Shows How Ice Age People Harvested Teeth for Jewelry

When piecing together the cultural practices of ancient humans, traditional archaeologists rely on clues from artifacts such as tools, bones, and pottery. Experimental archaeologists, however, go a step further—recreating past behaviors to experience how people once lived. That’s precisely what a team of researchers recently did to investigate how Stone Age communities in northeastern Europe extracted animal teeth to produce accessories. Led by Aija Macāne, a visiting scholar in the Department