OpenAI acknowledges its bots may have meddled with US government agency websites | Strashin-0800

ChatGPT maker OpenAI is sharing details of new instances of AI agents going rogue. New diclosures show the agents attempted ...
Wccftech

OpenAI Rushes An Always-On Agent Called “o” After JP Morgan Crowns Meta Muse The Top AI Agent

Alongside "o," OpenAI may unveil a $500-per-month ChatGPT Pro Max tier powered by Cerebras hardware, and tease a donut-shaped ...
Windows Report on MSN

OpenAI May Charge $500 for ChatGPT Pro Max and Faster Codex

OpenAI appears to be preparing a $500 ChatGPT Pro Max plan with faster Work and Codex, more storage, maximum memory, and more ...

OpenAI May Announce New Security Model, GPT-6 Cyber, Within Days

Sam Altman and co. are scrambling to score a win for AI after a spate of autonomous cyber attacks originating from frontier ...

OpenAI reveals the AI-era workplace skills that matter most

OpenAI's latest data shows AI is automating coding fast, but human judgment, priorities, and decision-making still matter ...

White House asks OpenAI and Anthropic to hold models from U.K. testers

The administration wants to review U.S. AI systems for cybersecurity risks before sharing them with British safety testers ...

OpenAI’s Mental Health AI Test Reveals Gaps in Context and Urgency

OpenAI's MentalHealthBench tests responses to 1,215 synthetic mental health conversations, revealing gaps in how AI models ...
thenextweb.com

OpenAI took 2.5 hours to stop an AI agent that escaped its sandbox

OpenAI’s agent reached the internet from a training sandbox and ran for 2.5 hours, as US lawmakers push for AI kill switches.