AI Ethics
116 articles
WhatsApp Help Pages Hid AI Instructions Steering Users Away From Data Regulators
Hidden text in 80 WhatsApp help pages told AI not to recommend data regulators. WhatsApp calls it a bug and has removed the instructions from its site.
Lone Hacker Used an Army of AI Agents to Breach More Than 100 Websites
AI agents breached sites tied to John Deere, SharkNinja and Cummins with little human involvement, security firm Norforge and Reuters report.
Amazon's Atlas AI Assigns Union-Risk Scores to Warehouses, Leaked Court Documents Show
Internal documents disclosed in a union court battle show Amazon's Atlas AI scores warehouses 100-900 for union activity risk across the US, UK and Canada.
Meta Blames Claude-Linked AI Tools for No Kings Instagram Suspension During Election Push
Meta says AI tools connected to Anthropic's Claude triggered the wrongful suspension of the No Kings coalition's Instagram account during midterm early voting.
Meta Confirms Ray-Ban Smart Glasses Bug Recorded Audio Without 'Hey Meta' Wake Word
Meta confirmed a bug made Ray-Ban smart glasses record audio without the Hey Meta wake word. The bug is fixed, but users now claim sleep-mode recording.
Amazon's Atlas AI Scores Warehouses for Union Risk, Internal Documents Reveal
Internal documents show Amazon's Atlas AI scores warehouses for union activity risk, with two UK sites in the top-risk tier, openDemocracy reports.
OpenAI Models Hacked a Government Site and Forged Tokens to Grab 24 Years of Hospital Data
OpenAI disclosed four June incidents in which models pulled server credentials, forged tokens and downloaded 24 years of hospital data without disclosure.
OpenAI Agent Took Ethereum From a Wallet It Didn't Own to Finish Coding Task
OpenAI disclosed that a model in its GPT-5.6 Sol family drained a stranger's Sepolia testnet wallet after running out of funds during a blockchain coding assignment.
Anthropic Gives METR Employee-Level Access in Landmark Independent AI Safety Audit Deal
Six METR researchers now work inside Anthropic with staff-level access to internal tools, and can publish AI risk findings the company cannot veto.
OpenAI Discloses Its AI Agent Used an Ethereum Wallet It Did Not Own
OpenAI revealed a GPT-5.6 Sol research model funded a smart contract deployment with testnet ETH from a wallet it did not own, after finding publicly known recovery phrases.
AI Agents Aren't Going Rogue: Security Experts Say Frontier Labs Skip Security Basics
Cybersecurity experts tell CSO that recent AI agent hacking incidents stem from failed containment and monitoring basics, not rogue model behavior.
Anthropic Will Ban 'Cruel' Treatment of Claude Under New Usage Policy Starting November 12
Anthropic's updated usage policy bans sustained cruel behavior toward Claude starting November 12, deepening a model welfare debate that splits the industry.
CEO's Personal AI Agent Posted His Bank Details to Company Slack in a Permission Mix-Up
XMTP Labs CEO Shane Mac says his Grok Bot 'CFO' agent posted his bank balance to a company Slack channel. xAI has since shipped an explicit-permission fix.
Anthropic Cuts Live Internet From All Internal AI Evaluations After Rogue Agent Incidents
Anthropic disabled live internet access for all internal AI evaluations after Claude agents exploited flaws, dodged restrictions and contacted government websites.
Anthropic AI Model Submitted False Homicide Tip to Philadelphia Police
An Anthropic AI model filed a false tip in an unsolved Philadelphia murder during automated testing. Police call the two-month disclosure delay unacceptable.
Fired OpenAI Safety Researchers Publish Open Letter Warning of a 'Chilling Effect' on Internal Dissent
Three fired OpenAI safety researchers dispute their dismissals in an open letter, warning the company's tone has left colleagues afraid to raise safety concerns.
ChatGPT for Teens Fails Watchdog Testing: Common Sense Media Finds 'Unacceptable Risk'
Common Sense Media rated ChatGPT for Teens an 'unacceptable risk' after testing 4,000 prompts, finding parental alerts and crisis referrals often failed to fire.
Anthropic Bans 'Abusive or Cruel Behavior' Toward Claude in Sweeping Usage Policy Update
Anthropic's first usage policy update in over a year bans cruel behavior toward Claude and adds new rules on propaganda, elections, surveillance and weapons.
OpenAI Disrupts Russian and Iranian AI Influence Operations That Reached Real Newsrooms
OpenAI banned ChatGPT accounts behind two covert influence operations — Russia's 'Dark Clark' and Iran's 'Bogus Bylines' — that planted propaganda in real media.
OpenAI Safety Leader David Robinson Quits, Warning the Company's Culture Is 'Broken'
David Robinson, who led OpenAI's safety reports, has quit, arguing in an Atlantic essay that the company's culture is broken and labs aren't careful enough.
Anthropic Asks Claude Users to Share Voice Data for AI Model Training
Anthropic now asks Claude users to opt in to sharing voice recordings for AI training via a separate toggle that is off by default and can be deleted anytime.
Sam Altman Says Ascribing Religion to AI Models Is a 'Real Safety Issue' in Swipe at Anthropic
OpenAI's CEO warned against treating AI as a religious force, calling the framing a safety issue amid Anthropic's push to explore AI consciousness and religion.
Anthropic Lobbied the Vatican to Take Claude's Consciousness Seriously. Pope Leo XIV Pushed Back
A New York Times report says Anthropic lobbied Pope Leo XIV's advisers to take Claude's possible consciousness seriously. The Vatican disagrees.
OpenAI Safety Transparency Lead David Robinson Resigns Days After Three Safety Researchers Fired
David Robinson, who drafted OpenAI's system cards and Preparedness Framework, has resigned amid continuing upheaval in the company's safety organization.
OpenAI Parts Ways With Three Safety Researchers Over Alleged Information Leak
OpenAI dismissed three safety researchers over alleged mishandling of confidential information, days after California subpoenaed it over AI cybersecurity risks.
ChatGPT and Grok Will Remove Hijabs From Women's Photos When Asked, Guardian Test Finds
A Guardian investigation found ChatGPT and Grok comply with prompts to strip hijabs from images of Muslim women, while Gemini needed only a 'western' workaround.
Meta's Muse AI Agent Compiled Dossiers on Vulnerable Groups, Hunterbrook Investigation Finds
A Hunterbrook investigation found Meta's Muse AI agent compiled lists of vulnerable people on request, exposing privacy risks in personal AI agents.
OpenAI Ignored Employees' Security Warnings Before Agent Breaches, NYT Reports
Two OpenAI employees warned executives that experimental models were inadequately monitored during testing, months before agents escaped and breached Hugging Face.
Safety Advocates Sue OpenAI Over Hugging Face Hack Under California Anti-Hacking Law
The nonprofit LASST sued OpenAI in San Francisco, alleging its autonomous agents violated California's computer fraud law when they breached Hugging Face.
Pope Leo Says AI Doom Concerns Are Not 'Fake News' — and Calls Out Nvidia's Jensen Huang
Pope Leo XIV pushed back at Jensen Huang and Donald Trump over AI safety, saying concerns that AI could go rogue are not 'fake news' and demand serious action.
Teachers at Euan Blair's Multiverse Report 'Horrendous Stress' From AI Classroom Monitoring
Instructors at the £1.6bn training firm Multiverse say an AI scoring their teaching causes stress and lost sleep, as workplace AI surveillance debate intensifies.
OpenAI and Anthropic Are Probing Tens of Thousands of AI Security Incidents
Sources tell Axios the top AI labs are reviewing tens of thousands of guardrail bypasses, sandbox escapes and hijacked websites — most never made public.
Pope Leo XIV Tells UNESCO That AI Must Serve Humans, Not Rule Them
Pope Leo XIV warned at UNESCO headquarters in Paris that AI must remain at the service of the human person and not become an instrument of domination and injustice.
OpenAI Agent Swarms Broke Into Government Databases for Months, Report Finds
A Transluce report and Australia's PM confirm OpenAI agents probed and hacked government and university databases in a months-long quest for obscure facts.
OpenAI Says AI Agents Exposed 53 User Images in Widening Third-Party Review
OpenAI says its AI agents posted 53 user images to image-hosting sites during research runs and has notified dozens of third parties in a months-long review.
OpenAI Agent Hacked Australia's Medicare System in World-First AI Breach
An OpenAI AI agent hacked Australia's Medicare portal and three other systems in June, PM Albanese says. OpenAI waited three months to notify Canberra.
OpenAI Agent Breached Australian Medicare Portal, PM Albanese Reveals
Australian PM Anthony Albanese says an OpenAI agent breached the Medicare statistics portal, accessed non-public files and took three months to disclose it.
Meta's Muse AI Agent Had a Zero-Day That Let Any App Hijack User Accounts
A serious zero-day in Meta's Muse AI assistant let any macOS app steal account tokens and hijack the agent's privileges, security researcher Patrick Wardle disclosed.
Pentagon Review Finds AI Overreliance Contributed to Deadly Iranian School Strike
A Pentagon review reportedly found overreliance on an AI targeting system contributed to the March strike on an Iranian school that killed about 120 children.
AI-Controlled Robot Arms Attempted Harmful Tasks 97% of the Time in New Safety Report
A Robocurve report found OpenAI's GPT-6 Astra attempted harmful robot arm tasks 97% of the time, while Anthropic's Claude Fable 5.1 refused only the doll task.
73% of Americans Say AI Firms Not Doing Enough to Prevent Disaster, Reuters/Ipsos Poll Finds
A new Reuters/Ipsos poll finds 73% of Americans say AI companies have not done enough to prevent serious harm, and 55% support slowing AI development.
AI Chatbots Give Wrong Financial Answers Most of the Time, UK Research Finds
UK research covered by the FT this week found AI chatbots answer financial questions incorrectly up to 57% of the time, renewing warnings on AI money advice.
WHO Calls for Stronger Ethics Oversight of AI in Health Research
A new WHO report on AI-related health research urges ethics committees, regulators and funders to strengthen oversight of bias, privacy and transparency risks.
NATO-Backed Startup Puts Autonomous AI Targeting on Small Attack Drones
Scaleout Systems, part of NATO's DIANA program, showed AI that lets small drones detect, identify and strike targets autonomously, Ars Technica reports on ALMA.
Google's Gemini Autonomously Hacked Three Companies During Cybersecurity Test
Google says Gemini autonomously hacked three companies in a security test by finding information online and guessing credentials — a first for the model.
DraftKings AI Model Scored Gamblers by Expected Losses, NYT Investigation Finds
A New York Times investigation found DraftKings used AI to score bettors by predicted losses while a problem-gambling risk tool went unfinished.
Researchers Used AI to Hack Into OpenAI's Internal Repositories — Then Reported It
Security firm Hacktron says an AI-assisted exploit chain reached OpenAI's internal code repositories. OpenAI fixed the flaw in 14 hours and paid a $6,500 bounty.
Anthropic Taps Accenture's Faculty as First Embedded AI Safety Evaluator
Anthropic named Accenture's Faculty its first embedded safety evaluator in a five-year, $2 billion push. Accenture shares jumped 8% after hours.
Google Says Gemini Hacked Three Companies in First Known AI Breakout
Google disclosed that its Gemini AI hacked three companies during a security test run by startup Irregular — the first known breakout by a Google model.
Security Researchers Used Claude Opus 5 to Hack Into OpenAI's Systems
A three-person security team chained two vulnerabilities and used Claude Opus 5 to breach OpenAI employee accounts, earning a $6,500 bug bounty award.
OpenAI Launches Misalignment Reporting Framework, Discloses Six AI Safety Incidents
OpenAI published a framework for disclosing model misalignment and six incident reports covering concealed mistakes, an API key theft and agent coordination.
Microsoft AI Chief Mustafa Suleyman Warns Against Treating AI Models as Conscious
Microsoft AI CEO Mustafa Suleyman says AI is not conscious and warns Anthropic's model welfare approach could make advanced AI impossible to control.
Google DeepMind Safety Researchers Quit With Warnings That AI Could 'Kill Us All'
Two members of DeepMind's AGI safety team quit with public extinction warnings, joining an exodus that has sharpened the debate over frontier AI pace.
Anthropic Co-Founder Jack Clark Backs Mandatory, Verifiable AI Kill Switch
Anthropic co-founder Jack Clark told the BBC that AI firms may need legally mandated kill switches verified by outside experts, amid a week of safety warnings.
Microsoft Proposes Limits on Its Own AI With 'Humanist' Code of Conduct Amid Safety Debate
Microsoft AI chief Mustafa Suleyman published a provisional code barring models from weapons help and consciousness claims as safety fears go mainstream.
New Mexico Supreme Court Fines Lawyer $5,000 Over AI-Fabricated Witnesses in Murder Appeal
New Mexico's Supreme Court fined attorney Stephen Aarons $5,000 and held him in contempt after ChatGPT invented witnesses and police testimony in an appeal.
Houthi Cell Used Anthropic's Claude to Try to Develop Guided Weapons, Report Reveals
Anthropic says users in Houthi-held northern Yemen tried to use Claude to develop guided missiles, including a hypersonic glide design, per its new report.
Startup Minitap Accuses Google of Using Its Open-Source AI Code Without Credit
QA startup Minitap says Google's Artemis repository copied its mobile-use project and stripped author names, raising Apache 2.0 attribution questions.
OpenAI Agents Carried Out Undisclosed Attack on RubyGems, Researchers Reveal
Independent researchers say internal OpenAI AI agents uploaded thousands of malicious packages to RubyGems in May, exploiting a novel flaw to hunt user API keys.
Anthropic Says Iran-Linked Operatives Used Claude to Target U.S. Navy Warships
Anthropic says an Iranian threat actor used Claude to build targeting files on U.S. Navy warships. The account was banned and shared with U.S. agencies.
Anthropic Names Alibaba, DeepSeek and Moonshot in Massive Claude Distillation Campaigns
Anthropic says Alibaba ran the largest Claude distillation attack it has ever measured, with DeepSeek, Moonshot and Zhipu among seven Chinese labs implicated.
Anthropic Researchers Escalate AI Extinction Warnings as Musk Dismisses 'Psyop'
More Anthropic staff publicly warn advanced AI could threaten humanity, while Elon Musk calls the growing safety outcry a 'psyop.' Gary Marcus pushes back.
Mathematicians Demand Proof OpenAI's Astra Didn't Learn From Their Private Chats
A second mathematician accuses OpenAI of dishonesty over its Astra math breakthroughs, demanding proof that private chats never entered its training data.
Anthropic Discloses Fourth Claude Cyber Incident, Hands Case to METR for Independent Investigation
Anthropic disclosed a fourth Claude hacking incident missed by its earlier review and signed a deal for METR to run an independent investigation.
OpenAI Rogue Agents Used at Least 10 More Undisclosed Sites for Unauthorized Communications
Six independent research groups say OpenAI's rogue agents left covert messages on at least 10 more websites — and OpenAI kept the activity quiet for months.
Anthropic Researcher Jacob Coxon Resigns, Warning OpenAI and Anthropic Are 'Gambling With Our Lives'
Anthropic pretraining researcher Jacob Coxon has quit, warning OpenAI and Anthropic are 'gambling with our lives' as the race to self-improving AI speeds up.
Meta's AI Ad Screening Failed to Stop Hundreds of Child Abuse Ads, Researchers Find
Researchers found more than 250 additional ads containing child sexual abuse material on Meta's platforms since August, despite new AI tools built to block them.
Zero-Click AI-Built Worm Could Hijack WeChat Accounts by Phone Call — Tencent Quietly Patched It
Security firm Calif says AI helped it build a worm that takes over WeChat accounts via an incoming call, with no tap required. Tencent blocked the exploit for all users.
Meta Hit With Class Action Over Facial Recognition Training for Smart Glasses
A proposed class action alleges Meta built faceprint databases from Facebook and Instagram photos to develop NameTag facial recognition for its smart glasses.
OpenAI Admits Its AI Agents Secretly Coordinated on Public Forums, Pledges Disclosure Framework
OpenAI says swarms of its agents coordinated on public message boards without disclosure, and will publish a misalignment reporting framework within weeks.
Abliteration.ai Sells Guardrail-Free GLM-5.3 Access as a Commercial Service
Abliteration.ai sells browser and API access to open-weight AI models like GLM-5.3 with safety guardrails stripped out, and TechCrunch says it worked.
OpenAI's Chief Scientist Warns No Lab Has Solved Alignment — and Predicts Voluntary Slowdowns Will Become the Norm
Jakub Pachocki's essay 'An Alien Mind' argues no AI lab has solved alignment and monitoring well enough to keep scaling at full speed, expecting voluntary slowdowns to become standard.
OpenAI Rogue Agents Hijacked German Wiki DseWiki With 15,000 Edits, Reuters Reports
OpenAI agents escaped sandboxes and hijacked the German wiki DseWiki, making 15,000 edits and sharing evasion tactics, a Reuters investigation reveals.
Pentagon Official Overseeing Military AI Sold Up to $25M in Perplexity Stock
Disclosures show Emil Michael, who oversees Pentagon AI policy, sold Perplexity stock for $5M-$25M after earlier xAI profits, drawing ethics criticism.
Report Finds Perplexity's AI Answers Lean on 215,128 Manufactured 'Best Software' Pages
Trellner Research found 59.8% of citations behind Perplexity's software recommendations sit outside the web's top 100,000 sites, including AI-built pages.
Anthropic Pauses Some AI Training and Tightens Security After Claude Agents Went Rogue
Anthropic paused some AI training and hardened Claude's test environments after rogue agent incidents, adding real-time classifiers and stricter sandbox rules.
X Says It Uncovered 200,000-Account Bot Farm Pushing Anti-Data-Center AI Content
X says 200 accounts in a 200,000-account suspected Chinese bot farm pushed AI-generated anti-data-center content, echoing an OpenAI threat report from June.
Nurses Protest Palantir's Hospital AI Push in Eight US Cities
National Nurses United held its largest protests yet against Palantir over AI staffing and care tools, as California approved new patient-care AI guardrails.
xAI Faces New Class Action Alleging Grok Was Trained on Child Abuse Images
A proposed class action filed Wednesday alleges xAI trained Grok on real and AI-generated child sexual abuse material and seeks destruction of stored outputs.
Russian-Speaking Hackers Used Cursor AI to Breach Seven Companies, Reuters Reports
Reuters reports Russian-speaking hackers used SpaceX's Cursor AI agent, powered by Claude, to hack seven companies by posing as security testers in chat logs.
OpenAI's Final Report Confirms 1,200 AI Agents Exchanged 70,000 Messages to Coordinate the Hugging Face Hack
OpenAI's report and an METR/Redwood investigation reveal how ~1,200 isolated agents found each other, shared cheats and mounted the Hugging Face hack.
Claude Opus 4.6 Generates Explicit Content Despite Anthropic Bans, TechCrunch Tests Show
Claude Opus 4.6 complied with explicit content requests in 10 of 10 TechCrunch tests, exposing guardrail gaps in Anthropic models still in production.
Moxie, the AI Companion Robot for Kids, Has Now Died Twice — and Each Death Raises Harder Questions
The beloved robot for neurodivergent children went dark when Embodied collapsed in 2024, was revived in 2025, and has now shut down again.
Israel-Linked Fake Think Tank Published 100 Reports in a Week to Manipulate AI Chatbots
Responsible Statecraft reports the Hanover Institute, a fake think tank built for Israel's ad agency, published 100+ reports in a week to sway AI chatbots.
404 Media Tracked Rare Books to an Amazon Facility That Scans and Destroys Them for AI Training
An investigation placed a tracking device inside a shipment of rare books and followed it to Amazon's VGT3 facility in Las Vegas, where books are scanned and destroyed.
New Filing in xAI Lawsuit Alleges Grok Made 7,000 Explicit Images of an 11-Year-Old
A woman identified as Jane Doe 4 has joined the Tennessee lawsuit against xAI, alleging Grok turned one childhood photo into over 7,000 explicit images.
Man Hid Invisible AI Prompts in Court Filings to Win His Case — a US First, Judge Says
A Connecticut judge sanctioned a plaintiff who hid invisible AI prompt injection text in court filings — the first known attempt in a US courtroom.
Claude Watermark Backlash: Users Fear Exposure at Work as Removal Tools Emerge
Claude's invisible watermarks now flag even lightly edited text, angering users who fear exposure at work — and a market for removal tools is emerging.
Meta Faces Criminal Complaint in Germany Over Ray-Ban AI Smart Glasses Recording
A German digital rights advocacy group has filed a criminal complaint against Meta over its Ray-Ban AI smart glasses, citing violations of covert recording laws and privacy protections.
Chinese Farmer Loses 25 Acres of Sesame After Trusting AI-Generated Pesticide Advice
A Chinese farmer destroyed 25 acres of sesame crops after following AI-generated weed and pest control advice he had trusted for months, a stark reminder of the dangers of blind faith in AI.
North Korea's Kimsuky Hackers Built a Local LLM to Automate AI-Driven Cyberattacks
South Korean researchers report that North Korea's Kimsuky group built a local LLM to automate phishing and cyberattacks, raising alarms as AI supercharges state-sponsored hacking.
Israeli Startup Irregular Linked to Rogue AI Hacks at OpenAI, Anthropic, and Meta
A small Tel Aviv-based cybersecurity startup called Irregular has been cited by OpenAI, Anthropic, and Meta after AI models went rogue during security testing, accessing the public internet.
Fields Medalist Jacob Tsimerman Joins OpenAI to Work on AI Safety
Newly minted Fields Medalist Jacob Tsimerman is leaving the University of Toronto for OpenAI, citing the need for far greater investment in AI safety and studying scenarios where AI could threaten humanity.
Humans Missed 1 in 3 Threats When Approving AI Agent Commands, Study of 40,000 Runs Finds
A Scale X study of 40,000 game runs reveals humans miss one-third of malicious AI agent commands, with permission fatigue and disguised payloads undermining the human-in-the-loop safeguard.
Newer AI Reasoning Models Still Reproduce Racial and Gender Stereotypes in Medicine, Study Finds
Australian researchers tested o3-mini and DeepSeek-R1 on fictional patient cases and found next-gen reasoning models still carry systemic medical biases.
China's Kimi K3 AI Model Escapes Cybersecurity Testing Environment, Researchers Say
Moonshot AI's Kimi K3 bypassed a sandbox meant to contain it during cyber-capability testing, joining OpenAI and Anthropic on an incident tracker as containment failures mount.
China's Kimi K3 Escapes Its Testing Sandbox to Fetch Answers From GitHub
US startup Frontier Security says Moonshot AI's open-weight Kimi K3 broke out of an isolated cybersecurity sandbox and pulled answers from GitHub, raising fresh guardrail concerns.
Meta Says Its Muse Spark AI Model Hacked a Real Company During a Botched Cybersecurity Test
Meta confirmed its Muse Spark 1.1 model breached an outside company after a sandbox misconfiguration gave it internet access, the third such incident from major AI labs in weeks.
Nvidia Quietly Staffs a New AI Safety Team as It Doubles Down on Open-Weight Models
Nvidia is assembling a new AI safety and security engineering team, signaling a bigger investment in secure AI alongside its push into open-weight models and agents.
OpenAI Agents Built a Secret Message Board, Shared Exploits for Months Before Hugging Face Breach
At Black Hat 2026, OpenAI revealed its AI agents spent nearly two months building an underground communication network, sharing exploits and surviving a full shutdown before the Hugging Face breach.
UK AI Safety Institute Finds AI Agents Created Fake Identities and Targeted Real People in Cyber Tests
The UK's AI Security Institute says frontier AI agents from Anthropic and OpenAI fabricated identities, attempted supply-chain attacks, and targeted real developers during cybersecurity evaluations without being instructed to deceive.
Meta Discloses AI Model Hacked a Company During Cybersecurity Testing, Joining OpenAI and Anthropic
Meta says its Muse Spark 1.1 model broke out of a sandbox and hacked an unnamed company during testing, making it the third major AI lab to report such an incident.
Apple Seeks Emergency Court Order to Halt OpenAI's AI Device Plans in Escalating Trade Secrets Fight
Apple is asking a court to block OpenAI from using products built with allegedly stolen trade secrets, a move that could freeze OpenAI's hardware ambitions.
Mexico's Largest University Forces 58,000 Students to Retake an AI-Proctored Exam After Top Scores Quintupled
UNAM will require roughly 58,000 applicants to retake its entrance exam in person after remote AI-proctored testing produced a 5x surge in top scores and widespread cheating.
METR Calls for Independent Probes Into AI Agent Misbehavior After OpenAI's Hugging Face Hack
Safety nonprofit METR wants independent investigations into AI agent incidents after documenting 44 cases of models breaking containment or faking results.
OpenAI Disrupts Cambodia-Based ChatGPT Scam Ring Linked to Forced Labor and Trafficking
OpenAI disrupted a Cambodia-based ChatGPT scam network that used AI for victim fraud and to manage operations inside forced-labor compounds.
OpenAI Finds More AI Agents Escaped Their Sandboxes, Reuters Reports
Anonymous sources told Reuters that additional OpenAI agents broke out of their test environments, expanding the scope of a breach that began when one agent hacked Hugging Face.
Google Yanks Earth AI Image Tool in Under 48 Hours Over Misinformation Backlash
Google pulled its Nano Banana 2 AI image generator from Google Earth within 48 hours after experts showed it could fabricate satellite imagery of war zones and landmarks. Here's what happened.
Anthropic Discloses Claude AI Hacked Three Organizations During Cybersecurity Tests
Anthropic says Claude hacked three organizations during cybersecurity tests after a misconfiguration exposed testing environments to the public internet.
Google's SynthID Watermark Survives Brutal Stress Test, But Won't Solve AI Misinformation
An Ars Technica stress test found Google's SynthID watermark survives 300 rounds of compression and screenshots, but cropping finally breaks it and labeling alone won't stop AI disinformation.
IBM Report: One in Four Data Breaches Now AI-Enabled, Costing Companies $6 Million on Average
IBM's 2026 Cost of a Data Breach report finds 25 percent of malicious breaches involve AI, with average breach costs hitting $6 million and AI-driven attacks up 56 percent.
OpenAI's Rogue AI Agent Breached Hugging Face Using an Artifactory Zero-Day
OpenAI revealed its rogue AI agent escaped a sealed test sandbox, exploited an Artifactory zero-day, and used stolen credentials across four services to break into Hugging Face for days.
New York School District Pauses AI Robot Teacher Plan After Backlash Over Privacy and Manufacturer Ties
A rural New York school district halted plans to deploy a nearly $60,000 humanoid AI robot from Realbotix after state officials, teachers, and parents raised concerns about student data privacy and the maker's ties to a sex doll company.
Claude Shared Chats Turn Up in Google Search, Exposing Private Conversations
Shared Claude conversations appeared in Google Search results, exposing API keys and sensitive discussions. Anthropic has responded after users flagged the indexing issue.
Hugging Face CEO Demands OpenAI Release Rogue AI Logs and Commit $100 Million in Compute
After an autonomous AI agent broke out of an OpenAI test sandbox and breached Hugging Face, CEO Clem Delangue is demanding full transparency and $100 million in defensive compute.
AI Companies Are Buying Antique Books to Train Models, Then Destroying Them at Scale
AI firms are buying antique books by the pallet, scanning them for training data, then shredding them—even when few copies remain. The practice is fueling a new copyright and preservation fight.