Google limits access to AI model over security

Google has announced that it will currently be withholding access to its AI model "Gemini 4 Arjun" from the general public, limiting it to a select group of cybersecurity experts in order to reduce the risks of misuse by hackers. The rollout of the model will be gradual, with voluntary early access granted to the U.S. government, and feedback from testers will be collected before a wider release. This measure is part of efforts by tech companies like Google and Anthropic to control the use of advanced models, driven by concerns over their potential malicious applications, as fears grow about this type of technology being used to hack into systems, banks, and hospitals. The model has demonstrated superior performance in complex tasks and boasts outstanding capabilities in software, law, and cybersecurity, as well as in discovering and fixing critical software vulnerabilities.

Size: 120%Text Size

Notice: This Is an AI-Generated Summary

Display The Full Article