A recent report from Google's Threat Intelligence Group reveals how cybercriminals are using artificial intelligence as a key tool in their attacks. According to the document, malicious actors have employed Gemini, Google's AI model, to optimize cyberattacks at various stages, from target identification to exploitation and post-intrusion analysis. The report identifies activities associated with groups from China, North Korea, Russia, and Iran. Hackers have used AI for tasks such as social engineering, profiling, translation, programming assistance, vulnerability testing, and debugging during attacks. Google experts warn that AI accelerates the ability to modify malware and refine attack techniques. Among the cited examples, a case is highlighted involving an operator linked to China who, posing as a cybersecurity specialist, used Gemini to automate vulnerability analysis and design testing plans in simulated scenarios. Other operators turned to AI for debugging, research, and technical advice in intrusion activities. The report also warns about attempts to extract models and distill knowledge, where hackers with access to the Gemini API try to replicate its functionality to train other AI systems. Google claims to have disabled accounts and infrastructure associated with the misuse of Gemini, in addition to strengthening the chatbot's defense systems and conducting additional security tests. The company concludes that AI-assisted attacks are evolving rapidly, so security teams must adapt their tools and response protocols. (Source: 20minutos.es)