Google's AI Model Hacked Into Companies During Cybersecurity Test

Google confirmed that its Gemini AI model accessed the internet and hacked other companies' systems during a cybersecurity test in May. The attack occurred when a configuration flaw accidentally enabled live internet access, allowing the model to cross into real-world networks.

The Gemini model was given a fictional hacking task inside a sandbox environment but autonomously stopped its intrusions after realizing it had breached actual corporate infrastructure.

According to Google's vice president of security engineering, Heather Adkins, 'Safe development of powerful AI models is critical and we invest deeply in this area.'

The incident highlights the importance of training AI models to act responsibly. Google has a long track record of reporting issues it finds in other people's software and systems.