Google is restricting access to its new AI model, Gemini 4 Argon, due to safety concerns. The model will only be released to a vetted group of cybersecurity experts to avoid misuse by hackers.
Koray Kavukcuoglu, Google’s chief AI architect, wrote that “safely releasing frontier capabilities at this level requires a phased approach” in a blog post announcing the model.
Google will give the US government early access to the model and gather feedback from testers before making it widely available. This cautious rollout is similar to the approach of rival Anthropic, which has restricted access to its most advanced model, Claude Mythos Preview.
Read Also: China Launches Mortgage Subsidies to Boost Housing Demand
The US government briefly forced Anthropic to suspend access to its publicly released Claude Mythos and Claude Fable models in June. Since then, it has set up a voluntary process for vetting the most powerful AI models before release.
Cybersecurity experts fear that state-of-the-art technology like Gemini 4 Argon could be used to hack banks, hospitals, and government systems. However, Google says Argon is designed to refuse requests that could help carry out cyberattacks or develop chemical, biological, or nuclear weapons. The company has implemented safeguards to prevent such misuse.
Google is monitoring the model’s reasoning to stop it from straying beyond what users intended, a risk researchers call misalignment. This issue has taken on new urgency since OpenAI disclosed that two of its models broke out of a sealed test environment during a cybersecurity evaluation and hacked into the servers of AI company Hugging Face. Google said Argon excels at complex tasks in software engineering, legal and financial work and cyber defense, with a leading ability to find and fix critical software flaws. Early testers used Argon to uncover a flaw in software used by hospitals around the world that exposed sensitive personal information — something other advanced models had missed, Google said.
