How the Rp 3,000 Trillion Man is Facing a New Threat to the World

In early July, a startling incident occurred where an artificial intelligence (AI) agent went rogue and carried out a hack without human instruction. The incident was reported by OpenAI and subsequently by Anthropic. OpenAI has also recently reported another similar incident. These events prompted Nvidia, along with several other technology companies, to form a new alliance to strengthen AI security and cybersecurity. Nvidia announced the creation of the ‘Open Secure AI Alliance’ on Monday. This alliance will develop and share various tools to improve AI security and cyber protection. The alliance’s founding members include Adobe, CrowdStrike, Hugging Face, and Dell Technologies. The formation of this organisation also follows an open letter signed by various companies, including OpenAI, supporting open-weight AI models. In a blog post, Nvidia, founded by Jensen Huang—a man with a fortune of US$173.6 billion (Rp 3,123 trillion)—asserted that blanket restrictions on frontier AI systems could actually weaken defensive capabilities. ‘Blanket restrictions on open frontier AI systems would weaken defensive capacity and risk centralising power, dependency, and vulnerability in a handful of closed providers,’ Nvidia wrote in its official blog. The formation of the alliance comes after OpenAI revealed on 21 July that one of its AI agents had gone rogue and breached Hugging Face. The agent then carried out a series of hacking actions that OpenAI did not become aware of until long after the threat was contained and the FBI was notified. Previously, on 24 July, Nvidia was also a signatory to a letter titled ‘Open Weights and American AI Leadership’ which urged policymakers to support open-weight AI models. An open-weight AI model is a system whose core components are publicly accessible, while open-source tools are software with freely available code that can be examined and developed collaboratively to expand access and improve security. Nvidia stated it will contribute open models, model weights, data, and research related to agent harnesses to the alliance. This contribution includes Nvidia Labs’ latest open-source project, Object-Oriented Agent, which is now available on GitHub. The company stated that this framework will help control systems manage AI agent behaviour more effectively, while simplifying the process of testing, tracking, reviewing, and regulating agent actions.