Govt Bars Officials From Uploading Classified Data to Public AI Tools

Government employees are currently prohibited from uploading classified materials to public artificial intelligence platforms due to apprehensions surrounding potential data breaches resulting from the dissemination of sensitive data across AI systems. The National CERT has released the National Cyber Security Handbook 2026-27, which outlines the guidelines of the Pakistan Information Security Framework 2026.

According to the new directives, public AI platforms are prohibited from being utilized for the transmission of official correspondence or software source code. Furthermore, citizens' personal information has been explicitly prohibited from being uploaded to public AI tools. The National CERT emphasized that sensitive data should not be input into AI platforms due to the risk of data leakage.

It instructed government officials to solely utilize department-approved AI tools. The handbook also instructs officials to eliminate names and sensitive details from AI prompts and files, and to promptly report any inadvertent disclosure of confidential information to the cybersecurity team. Government personnel have also been directed not to disclose passwords, administrative login credentials, or API keys to AI tools.

The guidelines also prohibit the installation of unauthorized AI extensions and plug-ins on official devices. All AI-generated content used in government operations must undergo scrutiny for accuracy and security. The National CERT clarified that AI-generated material cannot be employed in official matters without human oversight, and that the application of AI in government work will continue to be subject to security, privacy, and supervision prerequisites.