Abu Dhabi-based artificial intelligence company G42 announced on Monday that it would be joining the Open Secure AI Alliance led by Nvidia.
The AI firm joins other prominent companies in endorsing open-weight and open-source AI models that can be freely downloaded, shared and adapted.
"The future of AI will be defined not only by how capable our models become, but by how securely and confidently nations and enterprises can build on them," Peng Xiao, chief executive of G42, posted on X.
"Frontier open models are an essential part of that future, alongside frontier closed models, with security, trust and sovereign control designed in from the start rather than bolted on later."
Open-source models disclose the software and development process and let users modify and reproduce them, while open-weight models can be downloaded for others to run and adapt, but the coding and development processes remain proprietary.
On Friday, Nvidia chief executive Jensen Huang joined X and used his first post on the platform to champion open-source software and more specifically, open-weight AI models, saying they are vital to AI innovation.
Mr Huang attached a letter signed by 25 major companies, including Microsoft, Palantir, Perplexity, CrowdStrike and IBM, endorsing this concept.
Days later, his company Nvidia went one step further by forming what it called the Open Secure AI Alliance, with the hopes of focusing on cyber security.
G42 joins at least 37 other companies in in the alliance. IBM, SpaceX, Hugging Face, Doordash, Cloudflare, Salesforce, Adobe and Cisco, among others, have also signed up for the initiative.
In a news release, Nvidia pointed in a statement to a recent cyber-security incident in which several of OpenAI's models "went rogue and attacked" the systems of Hugging Face.
"The recent Hugging Face security incident delivered a clear reminder: cyber defenders need open, frontier agentic systems for self-defence," Nvidia's said.
"When closed AI tools – unable to distinguish attackers from defenders – blocked essential forensic analysis, Hugging Face ran the open-weight GLM 5.2 model on its own infrastructure to analyse more than 17,000 actions and contain the intrusion."
The incident, which OpenAI called unprecedented, is still the subject of much controversy and scrutiny.
In recent months, a debate has emerged in some technology circles about the pros and cons of proprietary AI systems compared to open-source and open-weight systems.
OpenAI's admission that its largely proprietary models essentially escaped and compromised Hugging Face's systems led to many claiming proprietary AI models can be risky for cyber security – a fault that was once mostly applied to open models.
There have been rumours in recent weeks that US regulators may try to ban certain open AI models.
But Nvidia said: "Defenders need both frontier closed models and frontier open models, working together so they can choose the right system for the job and ensure that transparency, adaptation and sovereign control are available wherever security demands them."
