Hacking groups funded by foreign actors and adversaries of the United States are increasingly using artificial intelligence to improve their odds of successful cyberattacks, according to analysis from some of the leading technology companies.

Microsoft and OpenAI released a report on Wednesday detailing how hackers from China, Iran, North Korea, and Russia are using large-language models to seek vulnerabilities in the software and security practices used by the U.S. government. The hackers are also using the technology to create “scripts” that can identify ways to break into government infrastructure that they can then use to steal valuable data and disrupt operations.