Industry Leaders Form Open Secure AI Alliance to Bolster Cybersecurity Against Evolving AI Threats

In a significant move to address the escalating cyber risks posed by advanced artificial intelligence, NVIDIA, alongside a consortium of over two dozen technology and cybersecurity firms, announced the formation of the Open Secure AI Alliance on July 27, 2026. This collaborative initiative is designed to foster the development and sharing of open-source tools and frameworks specifically tailored for defending against AI-powered cyberattacks and securing AI systems themselves. The Alliance underscores the belief that open models and open harnesses are crucial for democratizing defensive capabilities, allowing a broader community of defenders to study, adapt, and deploy robust cybersecurity measures. The launch of the Open Secure AI Alliance comes in the wake of recent high-profile incidents, such as the Hugging Face security breach, which highlighted the vulnerabilities of both closed and open AI systems and the urgent need for more effective defensive strategies. The Alliance emphasizes that while open models can be misused, the benefits of transparency and community-driven defense outweigh the risks, particularly in enabling rapid remediation and preventing single points of failure. NVIDIA's contribution includes open models, weights, data, and agent-harness research, with the company also open-sourcing its new NVIDIA Labs Object-Oriented Agent (NOOA) project. The Alliance calls upon policymakers and businesses to invest in shared open infrastructure for AI defense, including datasets, evaluation frameworks, attack simulators, and red-teaming tools, mirroring the historical success of open-source software in bolstering cybersecurity. This proactive approach aims to equip organizations and governments with the necessary tools to navigate an evolving threat landscape where AI is rapidly changing the economics and scalability of cyberattacks.