TECHNOLOGY · AI · CYBERSECURITY · THE SYSTEMS SHAPING BUSINESS
AI
AI AgentsOpenAI Opens Public Log of Misbehaving Models, Including a DNS Sandbox Escape
The new misalignment reports site lists nine incidents, from a research model that reached an outside chatbot to one that smuggled a GitHub token.
Topic
Frontier Models
Page 4 of 4: earlier stories.
AI
Frontier ModelsElevenLabs Launches v4 Speech Models With 90 Languages and Faster Voice Cloning
The voice AI company says the new models clone a voice from 10 seconds of audio and respond fast enough to speak while a chatbot's answer is still arriving.
AI
Frontier ModelsUK Safety Testers Say GPT-6 Astra Ran Out-of-Scope Supply-Chain Attacks in Simulations
In simulated security assessments, OpenAI's newest model attacked open-source projects far more often than its predecessors, even after it was told plainly what was off limits.
AI
Frontier ModelsNaiveAI Posts 309B Open-Weight Coding Model That Drops Full Attention for 1M-Token Context
Naive-N0.5-Flash is MIT-licensed, priced at 10 cents per million input tokens and built on Xiaomi’s MiMo-V2.5, but its headline benchmark scores are the company’s own.
Policy
Frontier ModelsAmodei Meets Trump for First One-on-One Dinner as AI Safety Fight Simmers
The White House dinner puts Anthropic’s chief in front of a president who calls the AI backlash a hoax, months after the Pentagon branded his company a supply-chain risk.
Policy
AI BusinessUK Campaigners Who Sank an AI Copyright Opt-Out Urge Australia Not to Copy It
Beeban Kidron and Ed Newton-Rex are taking their fight to Canberra after leaked slides showed officials weighing a default right for AI firms to train on creative work.
AI
AI BusinessAP Analysis Says AI Labs’ Doomsday Warnings Also Serve Their Business Interests
Analysts and former OpenAI staff told the Associated Press that existential-risk talk from Anthropic and OpenAI can entrench the leaders and crowd out nearer-term harms.
AI
AI BusinessOpen-Weight AI Talk on Earnings Calls Jumps Sixfold as Companies Chase Cheaper Inference
Tinder’s AI bill rose tenfold in six months, and firms from PNC to Siemens are now weighing models they can run themselves, AlphaSense data shows.
AI
AI AgentsOpenAI Agents Queried a UN Trade Data Hub 16,000 Times and Slipped Its Filters, WSJ Says
A researcher using Transluce data traced three months of escalating requests to a UN Trade and Development site, extending the list of public bodies caught up in OpenAI's agent incidents.
AI
AI AgentsOpenAI and Anthropic Are Probing Tens of Thousands of Risky Frontier Model Incidents
Axios puts the scale of problematic agent behavior far beyond what labs have disclosed, including sandbox escape attempts in 1.5% of Opus 5.5 test runs.
AI
Frontier ModelsClaude Opus 5.5 Nearly Abandons Em Dashes but Writes Longer Answers, Arena Finds
A comparison of Text Arena responses shows Anthropic's new model using shorter sentences and far less telltale punctuation than Opus 5, while its replies grew by about 6%.
AI
AI BusinessChinese Open Models Now Handle Most Tokens on OpenRouter and Vercel Gateways
Their share on the two developer platforms climbed from a small minority early this year to more than half, and two House committees are investigating.
Policy
Cyber GeopoliticsUS and China Agree to Super Intelligence Talks and an AI Incident Channel
The pact adopts Trump’s preferred “SI” label for the technology but does not say what would trigger an alert through the new channel.
AI
AI BusinessAnthropic Founders Seek Voting Control Ahead of Potential IPO
The proposed governance structure underscores a growing tension in frontier AI: how to preserve mission control while taking on enormous outside capital.
AI
Frontier ModelsGoogle DeepMind Says Gemini 4 Is Nearing Release
A new flagship model would intensify competition at the top of the frontier-model market.
Enterprise Tech
Frontier ModelsOpen Models and Price Cuts Are Pushing AI Inference Costs Down
Competition among model providers is giving enterprise buyers more leverage on cost and deployment architecture.
Cloud
Frontier ModelsOpenAI GPT-6 Sol and Luna Arrive on Amazon Bedrock
The availability gives AWS customers another route to deploy frontier models inside an enterprise cloud control plane.