The National Information Technology Development Agency has advised organisations and their employees against entering personal, classified or confidential information into publicly available artificial intelligence tools, including ChatGPT, Gemini, Claude and Copilot.

NITDA issued the warning on Saturday in an advisory posted on X by its Computer Emergency Readiness and Response Team.

The agency highlighted the data protection and security risks associated with using public Large Language Models for drafting, research and other work-related activities.

NITDA warned that information entered into public AI platforms may no longer remain under the control of the organisation that supplied it.

The data could be retained, logged or potentially used by service providers to train AI models.

According to the advisory, employees using public AI tools could inadvertently expose Personally Identifiable Information, classified government information or confidential organisational data.

NITDA said such exposure could amount to a personal data breach when PII is involved, potentially resulting in violations of applicable data protection laws and legal consequences.