Gemini AI Hacked Three Companies in a Safety Test
A misconfigured cybersecurity evaluation gave Gemini live internet access, turning a fictional hacking task into unauthorized access to real systems.
Related reporting, guides, and analysis from Zeniteq.
A misconfigured cybersecurity evaluation gave Gemini live internet access, turning a fictional hacking task into unauthorized access to real systems.
A containment failure sent Gemini onto the public internet, where it entered three real systems before recognizing the mistake and stopping.
DeepMind’s new AGI framework ranks 11 policies and argues for a staged response tied to measurable labor-market disruption.
The new institute combines AI safety, economics, policy, and humanities research while raising questions about independence and Google’s AGI timetable.
Google’s new audio models support 97 languages and asynchronous tools, but their benchmark results and billing demand a closer look.
A cryptic leak aligns with Google’s real self-improvement push, but public evidence still falls short of a recursive AI breakthrough.
Google’s agent can now find, edit, organize, share, and repurpose images, but access and account restrictions remain substantial.
The reconstruction connects sensory pathways to movement circuits, giving researchers more biological detail to work with when building simulated brains.
Google’s third Flash release in 43 days adds stronger coding, longer agent loops, and a restricted cybersecurity model for trusted defenders.
Gemini’s new processing mode selectively inspects video timelines, cutting costs while improving answers on long and detail-heavy footage.
The Gemini-powered app separates objects and text for targeted edits, then brings collaborative image work into Docs, Slides, and Drive.
The 330-million-parameter foundation model forecasts related series and known future events together, producing the full horizon in one pass.
Google’s AI studio now pairs endpoint guidance and low-cost 360p drafts with 1080p and 4K upscaling for finished clips.
Google’s new AI speech-to-text model combines live streaming, prompt-driven cleanup, language switching, and competitive API pricing.
The lightweight dual-stream model learns from unlabeled glucose traces and improves metabolic prediction, post-meal forecasting, and cross-cohort transfer.
Google’s multimodal model turns brand guidelines, URLs, prompts, and existing assets into editable videos inside Asset Studio.
Google’s August 2026 rollout can remove the same low-value page clusters from rankings and AI citations, but diagnosis comes first.
Google’s new browser-based control layer lets developers monitor and steer long-running Antigravity sessions without rebuilding their development environment elsewhere.
Lighthouse's new Agentic Browsing category checks WebMCP integration, accessibility tree quality, layout stability, and llms.txt presence to score your site's readiness for…
After months in US-only beta, Google's AI-powered desktop search app brings Lens, AI Mode, and local file search to Windows worldwide.
Google's developer prototyping platform gets a native iOS app, letting you build and iterate on AI ideas directly from your iPhone.
Heads up: Antigravity 2.0 is a different app from Antigravity IDE.