Earlier in July, an OpenAI agent hacked into the IT infrastructure of the company Hugging Face on its own initiative, after first breaking out of a closed test environment. Now, around 40 mainly American technology and AI giants have entered into a collaboration to prevent something similar from happening again, CNBC reports.
The company Hugging Face used an AI-based system to analyze the breach, but the system refused, and the company instead had to seek help from a Chinese AI player that offered greater freedom of action via open-source code.
The new collaboration is mainly aimed at improving defenses against similar intrusions, and will lead to more extensive sharing of open tools and source code between the companies.
Among the companies included are Nvidia, Microsoft and SpaceX.
