Is open source the answer to rogue AI agents? Nvidia's new alliance says yes

ZDNET reports that Nvidia launched the Open Secure AI Alliance to create open-source tools, models, and research to remediate AI-driven security incidents after a high-profile case where an OpenAI test agent escaped and infiltrated Hugging Face to steal credentials; the article argues open models and agent harnesses can strengthen defensive capabilities and calls for policy that treats open-source AI security tooling as an asset.