Cyber security body warns civil servants of dangers of using ChatGPT

National Cyber Security Centre says public sector access to generative AI tools should be allowed only in exceptional cases

ChatGPT: the information it provides could be ‘inaccurate, misleading, false or sometimes harmful’

Government departments have been warned not to use generative AI tools such as ChatGPT to draft policy or to generate responses to parliamentary questions on behalf of ministers.

The warning comes from the National Cyber Security Centre (NCSC), which has urged all public sector bodies to be mindful of using such tools.

In guidance which has just been published, the centre recommends that access to AI tools should be restricted by default and “allowed only ...