The US Department of Internal Security (DHS) recently conducted a pilot project aimed at identifying and eliminating vulnerabilities in critical government systems using artificial intelligence (AI). From the end of 2023 to the beginning of 2024, the Cybersecurity and Infrastructure Security Agency (CISA) team conducted an experiment to assess how much AI can increase the accuracy and speed of detecting vulnerabilities in networks and systems compared to traditional tools.
Two scenarios were utilized to test AI tools during the experiment: assessing the safety of networks of federal partners and testing in a controlled environment. Following the experiment, the CISA agency came to several conclusions:
- The best use of AI to detect vulnerabilities is to supplement and enhance existing tools, rather than replacing them;
- In some cases, training analysts to work with new AI capabilities is time-consuming, and the efficiency increase may be minimal;
- Sometimes, AI tools behave unpredictably, posing challenges in problem elimination.
AI tools are continuously being refined, and CISA plans to keep an eye on the market for new tools and conduct further testing. The pilot project underscores the importance of integrating AI into cybersecurity processes and demonstrates that even in its current stage of development, AI can significantly enhance the security of critical systems.