​FOI REF: FOI/2023/5046

You asked

​Have you since 01 January 2020 conducted any reviews into the use of AI? If so, please can you publish the findings along with methodology used?

We said

Thank you for your request. 

We have not conducted any reviews into the use of Artificial Intelligence (AI) since 01 January 2020. 

However, we have provided a small number of employees with ChatGPT accounts for the purposes of conducting an initial assessment of the service. This aims to cover a number of themes including: 

  • Considering the risks of using Large Language Models (LLMs) including in relation to data protection and security, and the presence of bias and inaccurate information 

  • Establishing business management and security operating procedures that would need to be applied if these tools were ever to be used 

  • Identifying potential use cases of the service 

This assessment is likely to take place over the course of several months.