AI
The latest in ai, ranked by importance and recency.
AxiosAnthropic Limits Access to Advanced Mythos Preview AI Model Due to Security Risks
Anthropic announced on Tuesday that it is withholding public release of its Mythos Preview AI model over concerns about its ability to detect and exploit vulnerabilities. The model is being provided to select tech and cybersecurity firms for defensive testing. This approach aims…
fortune.comMeta Unveils Muse Spark AI Model from Superintelligence Labs
Meta Platforms released Muse Spark, its first AI model developed under the Superintelligence Labs led by Alexandr Wang. The model powers Meta's AI products and shows performance improvements over prior Meta models, though it trails competitors in some benchmarks. Meta's stock ros…
EuronewsAnthropic Withholds Release of Advanced AI Model Mythos Due to Cybersecurity Risks
Anthropic announced its new AI model, Claude Mythos Preview, on Tuesday, stating it is too powerful for public release because of its ability to identify high-severity vulnerabilities in operating systems and browsers. The company is limiting access to select partners through Pro…
cnbc.comAnthropic Releases Preview of AI Model Mythos for Cybersecurity Initiative Amid Security Concerns
Anthropic has introduced a preview of its new AI model, Mythos, as part of Project Glasswing, a collaboration with over 40 companies including Microsoft, Amazon, Apple, and Google to enhance cybersecurity. The model has identified thousands of vulnerabilities in major operating s…
fortune.comAnthropic Withholds Release of Advanced AI Model Claude Mythos Due to Cybersecurity Risks
Anthropic announced its most capable AI model, Claude Mythos, on Tuesday, but decided against public release due to its potential to enable widespread cyberattacks. The model has identified thousands of previously undetected software vulnerabilities. The company is collaborating…
The New York TimesAnthropic Announces Claude Mythos Preview, Withholds Public Release Due to Cybersecurity Risks
Anthropic revealed its new AI model, Claude Mythos Preview, capable of identifying thousands of software vulnerabilities. The company is limiting access to a consortium of tech firms to prevent potential misuse in cyberattacks. This development raises concerns about AI's role in…
Magicpiano / Wikimedia (CC BY-SA 4.0)Anthropic Unveils Mythos AI Model Preview in Cybersecurity Initiative with Industry Partners
Anthropic announced a preview of its new AI model, Mythos, as part of Project Glasswing to counter AI-enabled cyber threats. The company is collaborating with over 40 organizations, including Apple and Google, to test the model's capabilities in identifying software vulnerabiliti…
WiredAnthropic Launches Project Glasswing with Mythos AI Model to Address Cybersecurity Vulnerabilities
Anthropic has introduced a preview version of its AI model, Claude Mythos, through Project Glasswing, a collaboration with over 40 tech and cybersecurity firms. The initiative aims to identify and mitigate vulnerabilities in software systems before broader release. Partners inclu…
The New York TimesFlorida Attorney General Launches Investigation into OpenAI Over FSU Shooting and Security Concerns
Florida Attorney General James Uthmeier announced an investigation into OpenAI and ChatGPT on Thursday. The probe examines potential harm to minors, national security risks, and a possible link to a 2023 shooting at Florida State University. The accused shooter reportedly consult…
Chris Martin (Flickr user cjmartin) / Wikimedia (CC BY 2.0)OpenAI Launches $100 Monthly ChatGPT Pro Subscription Tier
OpenAI introduced a new ChatGPT Pro subscription priced at $100 per month, targeting power users with expanded access to its AI coding assistant Codex. The tier provides five times more usage than the $20 Plus plan, enabling longer sessions. This follows user demand for intermedi…
cbsnews.com (News photo)Anthropic Introduces Mythos AI Model for Cybersecurity Applications with Limited Access
Anthropic announced its new AI model, Mythos, on Tuesday, focusing on cybersecurity defenses. The company is restricting full release and collaborating with over 40 organizations, including rivals like Apple and Google, through Project Glasswing. A preview version is being tested…
The AtlanticAnthropic Announces Claude Mythos Preview AI Model with Cybersecurity Capabilities
Anthropic revealed Claude Mythos Preview, an AI model capable of identifying thousands of cybersecurity vulnerabilities in major software systems. The model is available only to select tech partners including Apple, Microsoft, Google, and Nvidia for securing their products. Anthr…
Jordan McRae/Wikimedia Foundation and Midjourney AI / Wikimedia (Public domain)Meta Launches Muse Spark AI Model from Superintelligence Lab
Meta introduced Muse Spark, its first AI model developed by the Superintelligence Lab, which outperforms previous Meta models but trails competitors in coding tasks. The launch boosted the Meta AI app to No. 5 on the App Store from No. 57. The model includes features for analyzin…
WiredAnthropic Launches Project Glasswing AI Model for Cybersecurity Vulnerability Detection
Anthropic has introduced Project Glasswing, a new AI model designed to identify security vulnerabilities in software systems. The model, named Claude Mythos Preview, is being provided exclusively to select companies and organizations for defensive purposes. Partnerships include N…
Dietmar Rabich / Wikimedia (CC BY-SA 4.0)OpenAI Pauses UK Data Center Project Citing Energy Costs and Regulation
OpenAI has paused its planned data center investment in the UK due to high energy costs and regulatory issues. The project, known as Stargate UK, formed part of a larger UK-US tech investment agreement announced last September. This development affects Britain's strategy to expan…
Koreller / Wikimedia (CC0)Nature Publishes Two Studies on Multiomics, Deep Learning, and Satellite Analysis of Human Activity
Nature has released two separate research papers. One employs multiomics and deep learning to analyze regulatory syntax in human development. The other uses satellite imagery to examine volatility in human night-time activity patterns.
GB NewsLiverpool Suffers 2-0 Defeat to PSG in Champions League Semifinal First Leg
Liverpool lost 2-0 to Paris Saint-Germain in the first leg of their UEFA Champions League semifinal on April 30, 2025, at Parc des Princes. Goals from Desire Doue and Khvicha Kvaratskhelia secured the win for PSG. Manager Arne Slot kept Mohamed Salah on the bench, citing the need…
The DispatchTrump Backs Down from Threats Against Iran, Leading to Ceasefire Agreement
Former President Donald Trump has withdrawn from his escalated threats toward Iran, resulting in a ceasefire. This development cedes strategic leverage to Iran according to some analyses. The move occurs amid ongoing U.S.-Iran tensions and broader Middle East dynamics.
Dasfour2022 / Wikimedia (CC BY-SA 4.0)Federal Court Denies Anthropic's Motion on Supply Chain Risk Label in Defense Dispute
A federal court has denied Anthropic's motion to remove a supply chain risk label imposed by the Department of Defense. The ruling stems from Anthropic's legal battle with the Defense Department over artificial intelligence use in warfare. Separately, Trump officials may be encou…
Credit: Paolo Bovo, Army / Wikimedia (Public domain)US Army Develops Custom Chatbot for Soldiers Using Military Data
The US Army is creating an AI chatbot trained on real military data to provide soldiers with mission-critical information. This initiative reflects the Pentagon's increasing adoption of artificial intelligence technologies. Both Wired and Semafor reported on the development, high…
Benlisquare / Wikimedia (Public domain)Meta Releases Muse Spark AI Model from Superintelligence Lab
Meta announced Muse Spark, its first AI model developed by the Superintelligence Lab. The model showed improved performance over prior Meta AI systems but trailed competitors in coding tasks. One source reported the model offers health data analysis with associated privacy concer…
Photo: Zac Wolff / UnsplashRecent Surge in OpenAI News Developments Since Thursday
Multiple announcements and updates from OpenAI have emerged since Thursday, covering product launches, partnerships, and internal changes. These events include advancements in AI models and collaborations with tech firms. The volume of news has made it challenging for non-special…
IAEA Imagebank / Wikimedia (CC BY 2.0)OpenAI Releases Safety Blueprint Addressing Child Sexual Exploitation Amid Meta Legal Defeat
OpenAI has issued a new safety blueprint to combat the increasing incidence of child sexual exploitation online. This development coincides with Meta's loss in a multimillion-dollar lawsuit over failures to prevent child trafficking on Facebook and Instagram. The Guardian's 2021…
SpaceX / Wikimedia (CC0)SpaceX Plans IPO with Elon Musk Requiring Wall Street Firms to Subscribe to Grok AI
Elon Musk is conditioning Wall Street firms' participation in SpaceX's upcoming initial public offering on subscriptions to his Grok AI chatbot. The IPO could value the company at $1 trillion, positioning it among the largest in history. This requirement applies to advisory roles…
SemaforAnthropic Co-Founder Warns of Upcoming AI Capabilities for Exploiting Web Vulnerabilities
Anthropic's co-founder stated that powerful AI models capable of exploiting website vulnerabilities will emerge soon. The company's new model, Claude Mythos, identified unknown security flaws in major web browsers and operating systems. Financial authorities have responded by dis…
Los Angeles TimesGallup Poll Shows Increasing AI Use Among US Workers with Persistent Skepticism
A Gallup poll conducted in February indicates that more American workers are using artificial intelligence in their jobs, with about 3 in 10 using it frequently. However, skepticism remains common, with many non-users citing preferences for traditional methods, ethical concerns,…
Federal Bureau of Investigation / Wikimedia (Public domain)AI Assistant Poke Charges Billionaire $136,000 Monthly Fee
Poke, an AI assistant without a price ceiling, charged one billionaire $136,000 a month. Marvin von Hagen stated this pricing detail. The information highlights Poke's premium service model.
Brina Schenk with UBC and Douglas College / Wikimedia (CC BY 4.0)Microsoft Tests OpenClaw-Like AI Bots for Integration into 365 Copilot
Microsoft is testing AI bots similar to OpenClaw for its 365 Copilot assistant. The effort aims to enable autonomous operation around the clock to complete user tasks. The company plans to demonstrate features at its Build conference on June 2nd.
The Boston GlobeStudy Finds AI Chatbots Often Fail to Generate Accurate Differential Diagnoses for Symptoms
A study by Mass General Brigham researchers examined the performance of AI chatbots in producing differential diagnoses from patient symptoms. The chatbots failed to generate correct lists more than 80 percent of the time with basic information but improved to over 90 percent acc…
U.S. Air Force AFRL by David Dixon / Wikimedia (Public domain)Anthropic Report Warns AI Model Claude Mythos Preview Too Powerful for Public Release
Anthropic released a report on its AI model Claude Mythos Preview, stating it is too powerful for public release. The company expressed concerns about the model falling into the wrong hands. Matt Shumer, founder and CEO of a former AI company, discussed the report on CBS Mornings…
Unknown authorUnknown author / Wikimedia (Public domain)President Trump Posts AI Image Depicting Himself as Jesus, Draws Conservative Criticism
President Trump posted an AI-generated image on Truth Social showing himself as Jesus healing a sick person, surrounded by symbols including an American flag and the Statue of Liberty. The post has prompted criticism from some conservative supporters who described it as blasphemo…
Pseudopanax at English Wikipedia / Wikimedia (Public domain)Researcher Creates Fictional Illness and Fake Studies Funded by Foundations
A researcher has created a fictional illness and produced fake studies funded by the Professor Sideshow Bob Foundation and University of the Fellowship of the Ring and the Galactic Triad. Large language models warned people that the illness was real. @Nature reported on the event…
YandexART / Wikimedia (Public domain)Glossary of Common Terms in Artificial Intelligence
This article provides definitions for key terms associated with artificial intelligence. It covers concepts from large language models to hallucinations and other frequently used phrases. The glossary aims to clarify terminology as AI technology develops.
Wikideas1 / Wikimedia (CC0)AI Data Centers Expand Nationwide Amid Community Opposition
Tech companies are constructing over 4,000 AI data centers across the United States to support artificial intelligence development. Communities in areas like Archbald, Pennsylvania, have raised concerns about environmental effects, increased electric bills, and changes to local l…
Damian Putica / Wikimedia (Public domain)Generative AI and Its Reported Impacts on Art and Society in 2026
A 2026 article discusses generative AI's effects on the art world through reproduction technologies. It highlights criticisms including resource use and societal harms. Artists are noted as early predictors of these issues.
Another Believer / Wikimedia (CC BY-SA 3.0)OpenAI Releases Policy Paper on Industrial Policy for AI Era Amid Public Opinion Shifts
OpenAI announced a policy paper proposing ideas for a social contract centered on people-first approaches in the intelligence age. The document follows the company's acquisition of a podcast network and plans for a Washington DC office. This occurs as polls indicate rising public…
Kevin Payravi / Wikimedia (CC BY-SA 4.0)Robbie Williams States He Uses AI for Writing Assistance on Instagram
Robbie Williams, a 52-year-old singer and former Take That member, discussed his use of artificial intelligence in an interview. He explained that he drafts content and then inputs it into AI for corrections. The interview occurred as he marked 30 years as a solo artist.
Marie-Lan Nguyen / Wikimedia (CC BY 3.0)Runner Derek Wallbank Uses ChatGPT for Paris Marathon Preparation
Derek Wallbank employed ChatGPT to assist with training, diet, and discipline ahead of the Paris Marathon. His experience involved testing the AI's capabilities through trial and error. The process highlighted the AI's strengths and limitations in supporting athletic preparation.
Ars TechnicaAI Models from Major Companies Show Low Accuracy in Predicting Premier League Soccer Outcomes
A study reported by Ars Technica evaluated the performance of AI systems from Google, OpenAI, Anthropic, and xAI in predicting soccer match results in the English Premier League. The models, including xAI's Grok, achieved low success rates in betting simulations. The findings hig…
csmonitor.comAnthropic Releases AI Tool Capable of Detecting and Exploiting Software Vulnerabilities
Anthropic has introduced a new artificial intelligence tool named Mythos that identifies software flaws and demonstrates methods to exploit them. The tool's dual capabilities have prompted technology industry leaders to evaluate associated risks. This development occurs amid ongo…