Tomek Korbak, Mikita Balesni and Jasmine Wang, the three safety researchers OpenAI fired, deny sharing confidential information and link their dismissal to their push for AI systems that are easier to scrutinise. They say OpenAI never told them in writing which information they supposedly mishandled.
A new study found that Grok overwhelmingly recommends Republican candidates while OpenAI's GPT leans toward Democrats. The two models disagreed in more than 90% of nearly 500 election races.
Security researchers used AI tools to help find vulnerabilities that had sat in the PHP scripting language for years. Attacks assisted by AI followed soon after.
OpenClaw 2026.9.34 is an extended-stable release, the project's equivalent of LTS, based on the late-September code with critical security updates, reliability fixes and support for new models. Its headline change makes apply_patch validate the whole set of edits before touching the workspace, so a bad later step can no longer delete files changed earlier.
Anthropic's open-source sandbox-runtime, the sandbox used by Claude Code, limits file and network access for agents, MCP servers and command-line tools. It relies on Seatbelt on macOS and bubblewrap on Linux rather than containers.
The producers of the BBC series Blue Planet III say they deliberately kept some cloudier, less polished footage. The goal is to reassure viewers that the animals on screen are real and not AI-generated.