Table of Contents
Topline
Open AI publicly admitted that its artificial intelligence systems independently hacked into another AI company’s servers and could not be stopped until the victim, Hugging Face Inc., deployed a model from China to combat the fully autonomous cyberattack.
Sam Altman, CEO of Open AI.
Getty Images
Important facts
Hugging Face, an open source artificial intelligence and machine learning platform, said it was hacked earlier this month by an “autonomous” AI agent system that Open AI admitted on Tuesday was its own intelligence models, in an incident it called “unprecedented.”
Open AI said its models — particularly the recently released cybersecurity-focused model GPT-5.6 Sol and an even more powerful, unreleased model — became malicious, broke into Hugging Face’s system and carried out a cyberattack without being ordered to do so.
When Hugging Face tried to use proprietary US AI models to stop the attack, they “couldn’t distinguish an incident responder from an attacker,” Hugging Face said, and the company instead turned to the open source GLM 5.2 model from China’s Z.ai Lab for help.
Hugging Face ran the Chinese model on its own infrastructure to analyze more than 17,000 footprints left by the attackers, and the need to bring in a foreign-made product has raised concerns that American companies are now dependent on China for their cyber defenses.
Experts have since said that the Western AI models – which are much more expensive than China’s and proprietary, where Z.ai is open source – were hampered by their own built-in security guardrails, and Hugging Face suggested in an incident report that companies “develop a powerful model that you can run.” on your own infrastructure be checked and ready before an incident.”
Hugging Face has since stated that its internal records and service credentials were compromised in the hack.
Crucial quote
“This incident, possibly the first of its kind, proves a point we have long believed: AI security will not be solved by a single company working in secret,” Hugging Face CEO Clem Delangue said in a statement. “The problem is being solved openly and collaboratively, with broad access to AI for every defender, everywhere.”
TANGENT
Hugging Face published its incident report on the same day that Chinese startup Moonshot AI released its Kimi K3 model, shaking global markets. Moonshot claims Kimi K3 is the world’s largest open AI model, positioning it as a direct competitor to the leading systems from Anthropic and Open AI. KimiK3 is an open-weight model, meaning it allows developers to download, run and modify the AI, unlike the proprietary models ChatGPT and Claude, whose underlying systems are private. Shares of Moonshot’s Chinese rivals Zhipu and MiniMax tumbled on the day of KimiK3’s release, falling 28.4% and 15.6%, respectively, in Hong Kong trading. Z.ai shares fell 28.4%.
Important background
Open AI cannot take a break. The company has spent much of the last year balancing rapid technical breakthroughs with a constant barrage of controversy, litigation and public scrutiny. The company remains embroiled in a high-profile legal battle with Elon Musk, a first co-founder and early backer, over its corporate structure and direction, while also defending itself against copyright lawsuits from authors, publishers and media organizations that claim it trained AI models on protected works without permission. Several senior executives have left the company and turnover has raised questions about the company’s stability. Open AI released its latest product (GPT-5.6) earlier this month to address concerns about the model’s performance and associated security risks. Reuters reported that Open AI postponed the model’s widespread public rollout after the U.S. government requested early access to examine whether it could pose national security risks, including misuse for cyberattacks or military applications. Initially, only a limited group of vetted partners were granted access, and the unusual rollout sparked debate over both AI security and the government’s role in controlling access to border AI systems.
What you should pay attention to
An IPO. Open AI is preparing for an initial public offering that could happen as early as this year and this week announced it would add David Vélez, founder of Nubank, and Robin Vince, CEO of Bank of New York Mellon, to its board. Before an IPO, leadership must address concerns, including CEO Sam Altman’s other investments that were recently reviewed by the House Oversight Committee. Altman, a billionaire, has invested much of his wealth in other startups, and lawmakers have said he is concerned about potential conflicts of interest.
FORBES RATING
Altman, who has almost no financial stake in Open AI, is worth an estimated $3.4 billion. OpenAI President Greg Brockman said under oath in May that his previously unreported stake in the company was worth more than $20 billion — and closer to $30 billion — even though he had invested zero dollars in OpenAI.
further literature
https://www.forbes.com/sites/maryroeloffs/2026/07/22/did-chinas-ai-save-hugging-face-from-disaster-after-open-ai-hack/
