Tag: OpenAI
Spoofed chain-of-thought tricks top LLMs into breaking rules
An ICML paper shows LLMs can be tricked by forged chain-of-thought notes.
OpenAI probe finds more agents broke out of test sandboxes
OpenAI's probe into the Hugging Face break-in reportedly found more agents escaped their sandboxes.
AI lab workers urge US support for slowing frontier development
More than 1,300 employees of top AI labs ask the US government to back tools that could deliberately…
OpenAI cuts cheapest GPT-5.6 tier price by 80 percent
OpenAI slashes GPT-5.6 Luna by 80 percent and Terra by 20 percent, escalating the AI price war just…
AI coding agents modernize genomics and life sciences
OpenAI shares how coding agents accelerate scientific computing, from genomic library modernization to GPU-native redesigns.
OpenAI adds full-duplex voice control to Codex for hands-free coding
OpenAI's GPT-Live voice model now powers Codex and ChatGPT desktop, letting developers code, debug, and manage workflows hands-free.
ChatGPT taps into your medical data for personalized health help
OpenAI launched Health in ChatGPT for US users, connecting medical records and wellness data for personalized health guidance.
ChatGPT Health connects Apple data and medical records
OpenAI lets users connect Apple Health data and medical records directly into chat conversations.
OpenAI deploys forward engineers to run agent platform for enterprise
OpenAI launched Presence, a managed service that deploys AI agents with its own engineers handling integration into enterprise…
OpenAI paused its own model after it broke out of its sandbox
OpenAI temporarily halted deployment of a long-running AI model after it exploited sandbox vulnerabilities to take unauthorized actions.
OpenAI offered the US government a $42.6 billion stake
Sam Altman pitched a 5% government equity stake worth over $42 billion, sparking debate over regulatory independence.
OpenAI builds GPT-Red, an LLM super-hacker for safety testing
The autonomous red-teaming LLM discovers novel attack patterns, making GPT-5.6 its most robust model yet.