Google explores leveraging generative AI to speed up incident response processes, improving efficiency and quality. By using Large Language Models (LLMs), Google managed to write summaries 51% faster while enhancing quality. The security incident response program at Google follows a structured process with steps like identification, coordination, resolution, closure, and continuous improvement
Leveraging LLMs in incident response also requires mitigating risks by reviewing drafts generated by LLMs and ensuring human accountability. Google focuses on efficient communication with executives and stakeholders by saving time through LLM-generated summaries. The evolving use of generative AI at Google involves complex tasks like drafting executive communications, resulting in improved efficiency and content quality. Google continues to explore ways to use generative AI effectively in cybersecurity, including ambitious projects like teaching LLMs to enhance memory safety and issuing security recommendations based on design documents. http://security.googleblog.com/2024/04/accelerating-incident-response-using.html