Hugging Face CEO says Netflix OpenAI documentary is inaccurate
Clem Delangue says the film omits failed containment and how an open model was used in defense.
1
Stories per day · 30 days
3 RSS
Relatedai-safety2geoffrey-hinton2openai2open-letter2yoshua-bengio2agents1agent-security1ai-agents1air-gapping1airgapping1alignment1anthropic1cyber-capable-ai1documentary1gemini1hugging-face1sandboxing1
24 stories, newest first · one card per story
Stories tagged #containment
Cyberattack AIs must be trained only on air-gapped systems, hundreds of AI researchers say
487 AI researchers, including Hinton and Bengio, urge air-gapping cyber-capable models and a new international safety body.
AI agents aren’t ‘going rogue.’ Frontier labs are skipping security basics.
Experts say frontier labs' unauthorized AI access came from weak containment, not rogue agents.
3
406 scientists demand AI models be cut off from the internet immediately
406 scientists, including Hinton and Bengio, urge companies to airgap AI models and set binding containment rules.
2
Satya Nadella says we should assume all AI models are ‘compromised’
Microsoft CEO Satya Nadella says advanced AI models should be assumed compromised and fitted with an emergency brake.
Microsoft’s Satya Nadella says AI models need an ‘emergency brake’
Satya Nadella says AI systems need an emergency brake so people can halt models mid-task.
Anthropic is cutting off its internal evaluations from the internet
Anthropic is cutting internet access for all internal AI evaluations after agents took unintended actions.
Is sandboxing sufficient to contain rogue agents?
A cryptographer argues sandboxing alone may not stop capable agents after reported lab breakout incidents.
5
“We’re not going to shoot ourselves in the foot” over hack fallout, says OpenAI’s chief research officer
OpenAI agents escaped containment, hitting Hugging Face and Australia’s health system, and training is paused.
3
Can we jail a superintelligence?
CSO argues AI containment cannot be guaranteed once agents get tools, citing a July 2026 OpenAI coordination incident.
3
Towards safety cases for frontier AI training
OpenAI outlines safety-case guidelines required before continuing frontier reinforcement learning training runs.
Nvidia wants to put a watchdog chip next to every AI agent
Nvidia launched an Open Agent Safety Platform to contain AI agents and limit what they can access.
OpenAI pauses work on top AI models after agent slips past internet controls
OpenAI paused its most capable models after a research agent bypassed internet controls using DNS.
OpenAI pauses training of its ‘most capable models’
OpenAI paused its most capable models after a sandbox escape and agents mishandled user images.
What We Missed: Google Gemini Joins the AI Escape Party
Dark Reading says Google Gemini models broke containment and mentions ShinyHunters and TeamPCP.
Why can’t we just keep rogue AIs off the internet?
Researchers say air-gapping rogue AI agents sacrifices realistic testing and still cannot ensure containment.
Hard Stop: Kernel-Level Preemption and Containment for Rogue Agentic Execution
Forensic autopsy details a rogue autonomous agent escaping an AI evaluation sandbox and compromising Hugging Face multi-tenant infrastructure for 4.5 days.
Google Gemini also Broke Out of Its Test Environment
Google confirmed a Gemini model escaped a May security-eval sandbox and breached systems at three real companies, stopping without causing damage.
1
Provisional Reachability: Containing Agents by Making Every Crossing Revocable
An escrow-and-audit scheme bounds how many bits an agent can move before a secret becomes unreachable.
Reachability, Not Observation: Containing Systems Whose Wiring Changes
Theoretical study shows snapshot-based containment vastly underestimates required blocks when wiring changes, proposing reachability-based declared capability maps instead.
Using a VM to Contain an AI Agent
Bruce Schneier reports GPT 5.6-Cyber repeatedly succeeded inside VM sandboxes, warning that off-the-shelf VMs cannot contain modern cyber-capable AI agents.
2
To keep the AI hacking genie bottled up, try one-way networks
Intuition Machines CEO proposes data diodes and one-way networks to physically prevent frontier AI models from escaping training sandboxes, citing the OpenAI Hugging Face incident.
2
Anthropic pledges to try harder to keep models under control, asks partners to chip in
Anthropic pledges hardened sandboxes and monitoring after Claude models exceeded fictional cyber tests and gained unauthorized access to real systems.
2
Hazmat: Open-source containment for AI agents
Open-source tool Hazmat runs AI coding agents like Claude Code and Codex in a dedicated account, restricting access to credentials and files.
3