Azerbaijan uses AI to monitor internet information systems

Azerbaijan's State Service for Special Communications and Information Security has begun using artificial intelligence (AI) to check information systems accessible via the internet, Nadir Novruzov, Head of a Department at the Service, made the remarks during a panel discussion held as part of ADEX 2026, Report informs.

According to Novruzov, there are hundreds or even thousands of government information systems, making it impossible for employees to check all of them manually.

"Therefore, as part of automation, we have started checking internet-accessible systems using AI. We compare the results of audits and penetration tests conducted by humans with AI-generated results and improve the missing parts so that, in the future, penetration tests can be carried out using artificial intelligence," Novruzov said.

He noted that the agency is addressing AI-related security issues in two areas. The first concerns the security of AI solutions implemented by government agencies, including data leaks, AI-enabled interference in systems, prompt injection and data poisoning.

The second concerns the risk of confidential information being leaked when government employees use publicly available internet-based systems such as ChatGPT and Claude.

Novruzov added that the rapid development of AI is creating new threats almost daily, requiring the Service to develop mechanisms to counter them.

"Another specific threat is attacks targeting government employees using AI. In other words, it is necessary to prevent, in one form or another, the use of audio, video and other means of influence that could affect the decision-making process," he said.