Crowdstrike: These services aim to proactively identify and mitigate vulnerabilities in AI applications, including LLMs, by leveraging threat intelligence and simulating real-world attack scenarios.