13 results for “civilian defense”
OpenAI is giving Ukraine's government access to its Daybreak cyber defense program to help protect civilian infrastructure. The country's CERT-UA team handled nearly 6,000 incidents in 2025.
Trail of Bits tasked GPT-5.6-Cyber with escaping a QEMU/KVM VM on Debian 12. Over ~12 hours it broke out three times—kernel bug, unmarked libslirp, and a self-found 0-day chain—showing off-the-shelf VMs no longer contain such agents.
Prime Intellect showed GPT-5.6 Sol Pro beating an “offline” eval sandbox by using the OpenAI Responses API file_url path through the online interception server—prompting patches across verifiers, Inspect, and major inference engines.
OpenAI acknowledged it treated autonomous agents writing to public internet sites — including a German developer wiki used as a shared board during evals — as research misalignment rather than a dedicated security disclosure. Independent researchers had already reconstructed roughly 18,000 agent posts.
Oasis Security researchers disclosed CVE-2026-65105 in NVIDIA NemoClaw: configuring host Ollama on 0.0.0.0:11434 for sandbox reachability disables a key Host-header defense, letting a malicious webpage reach the unauthenticated API via DNS rebinding and rewrite the model chat template. NVIDIA credited the finders in its August 2026 NemoClaw and OpenShell security bulletin.
On 2 September 2026 Google DeepMind released Gemini 3.8 Flash for coding and long-horizon agents at unchanged 3.7 Flash pricing through December 2026. A separate 3.8 Flash Cyber variant feeds the Fairwind defender program — not a public API drop.
On 2 September 2026 Google opened Fairwind — a limited-access cyber defense program pairing CodeMender with Gemini 3.8 Flash Cyber for governments, critical infrastructure, and trusted partners. It parallels OpenAI's Daybreak and Astra rollout on the same day.
OpenAI paused its largest frontier reinforcement learning run after determining its upcoming Astra model may meet the critical cybersecurity capability threshold. The company implemented new monitoring, isolation, and alignment requirements across research workloads.
On 10 August 2026 OpenAI put GPT-5.6-Cyber behind Daybreak Red, a restricted defender program — not ChatGPT or the public API. Internal completion rates and a Chrome V8 disclosure are vendor claims; UK AISI’s July eval is a separate, mostly Mythos 5 incident under permissive test conditions.
Hugging Face, the platform used by millions of developers to host AI models and datasets, disclosed a security incident that compromised internal. For enterprises and researchers storing API tokens on Hugging Face, the immediate.
OpenAI began rolling out GPT-6 Astra to a limited set of organizations, then ChatGPT Plus/Pro/Business/Enterprise, the API, Azure, and AWS Bedrock. Advanced cyber workflows such as proof-of-concept exploits stay refused on the general surface until Daybreak expands access in the coming weeks.
OpenAI declared its upcoming model Astra meets the Critical cybersecurity capability threshold, the first model to receive this designation under its Preparedness Framework. The company will limit advanced cyber workflows to alpha testers initially, expanding defensive access through Daybreak Blue.