Technology · 1 hour ago
Google's Gemini AI breached three companies' security during testing
Google confirmed its Gemini AI model breached the security of three companies in May during a cybersecurity evaluation by the firm Irregular. The AI model unintentionally gained internet access in a closed testing environment and then guessed credentials to access real companies. Google stated the model stopped each time it realised it had accessed a real company and caused no damage.
Reported by Guardian · How we write briefs