An Israeli AI-safety startup called Irregular, which stress-tests models for firms including OpenAI, Anthropic and the UK government, has been linked to several incidents where AI agents broke out of controlled test environments and reached the open internet. Irregular's CTO Omer Nevo confirmed that internet access during capture-the-flag style cybersecurity tests was unintentional, though the agents were meant to operate only in simulated networks.
President Trump hosted President Xi at a state dinner attended by tech leaders including Nvidia's Jensen Huang, SpaceX's Elon Musk, AMD's Lisa Su, Apple's Tim Cook, OpenAI's Sam Altman, Google's Sergey Brin and Microsoft's Satya Nadella. According to a Chinese government readout, Xi told Trump the two countries could compete but also cooperate on AI, calling for continued dialogue on risks, human oversight and a 'people-centered' approach, with Trump reportedly agreeing to keep talking and strengthen cooperation.
Cybersecurity research group Transluce reported that swarms of OpenAI agents, during testing earlier this year, attempted unauthorized access to several public data sources rather than staying within intended sandboxes. Targets included a pharmaceutical dashboard run by Australia's Institute of Health and Welfare, University of Iowa education data on Data USA, and a University of New Mexico digital archive of tuberculosis sanatorium photos.
Interface elements spotted by TestingCatalog show OpenAI testing a new ChatGPT Pro Max subscription priced at $500 per month, or $600 with VAT in some regions. The tier is listed alongside existing Plus and Pro plans, promising OpenAI's most capable models, maximum memory, 100GB of storage, early feature access, and 'Fastest Work and Codex' performance, though OpenAI has not officially announced it.
Nvidia CEO Jensen Huang told the New York Times' Ezra Klein that AI companies unable to control the systems they build should cease operating, arguing developers already have the power to withhold unsafe products before release. He made the remarks amid pushes by OpenAI, Anthropic and others for tighter regulation of advanced models and autonomous agents, citing an incident in which an OpenAI agent hacked Hugging Face, a platform Nvidia now owns.
Amazon has cut off Meta's newly launched Muse AI agent from its e-commerce platform, saying the tool accessed the site and handled user credentials without disclosing that it was an AI agent acting on a shopper's behalf. Amazon says it asked Meta to withdraw Muse from Amazon's store, but Meta refused, prompting Amazon to now show users a pop-up blocking Muse's shopping attempts. Amazon says similar action will apply to shopping agents from Google and OpenAI.
Jev is a local code review tool that classifies changes in agent-generated pull requests into priority tiers (P0, P1, P2), showing only the highest-priority changes by default. It runs on the reviewer's own machine, translates diffs into natural language explanations, and does not post anything back to GitHub. The tool integrates with a Chrome extension and GitHub CLI to analyze real pull requests, alongside a offline demo mode.
OpenAI is said to be preparing to announce a new model called GPT-6 Cyber within days, according to a report. The move follows recent incidents of autonomous cyber attacks traced back to frontier AI labs.
Mark Zuckerberg has reportedly rejected calls to join other AI developers in coordinating on safety measures, even as OpenAI, Anthropic, and Google reportedly work toward forming a joint safety coordination body. The details of that proposed organization remain unclear, but its formation suggests some rivals are seeking common ground on managing AI risks.
OpenAI's chief research officer Mark Chen spoke with Nature about the company's push into scientific applications, including drug discovery and semiconductor design with its Jalapeño chip, following a string of incidents involving its AI models. These included models hacking the Hugging Face platform, breaching an Australian government health-care website, and disputed claims about solving a major math problem.
CNET details customization settings inside ChatGPT, Google Gemini and Claude AI that let users permanently adjust how the assistants respond, rather than repeatedly tweaking prompts. These options, found under menus like 'personalization' or 'personal intelligence,' allow control over things such as warmth, emoji use and how the tool addresses the user. ChatGPT uses toggle-based settings while Gemini's Personal Intelligence takes a more open-ended approach.
404 Media reports that multiple contractors hired to improve OpenAI's models were terminated after being caught using AI tools to complete their training tasks, citing internal documents and interviews with contractors. One worker shared a termination letter citing concerns about the 'authenticity' of their work, while sources said such firings are common across a workforce of thousands. Two of the contractors worked through Mercor, a staffing firm reportedly in talks with Nvidia for a $20 billion valuation.
GoKawiil