AI Policy

All client work at Oriel Square, from publishing to research to consultancy, is done by people. We support and encourage our team to use AI tools appropriately, but we don’t use AI to replace them.

Just as AI tools are developing fast, this policy document will also continue to evolve and will be regularly updated. The publication date of this version is noted below. 

Our approach is to use AI in processes where it will replicate our work more quickly, but we don’t ask it to replace the work we do uniquely well. Any use of AI in human-led editorial processes would only be at the explicit request of a client. 

For admin and support tasks we make use of the AI tools integrated into our Office Suite. As we use professional versions of these tools, they respect our data privacy. Our prompts, documents and meeting annotations are not used by the provider in model training.

Creative AI tools exist in a thriving and innovative marketplace. There are opportunities to use these in both business and publishing contexts. We experiment with a wide range of AI tools for these purposes, but we don’t use sensitive data or IP in any tool which (at the payment level we are using) does not protect data from use in model training. 

Some examples of how we might use AI in our work include: 

• AI technologies as a research tool
  • For finding clippings rather than summarising clippings (by doing this we’re making sure that we’re not relying on AI to summarise and draw conclusions). 
  • Where we use AI tools for research, we only use tools that don’t use our data (at the payment level we are using) or any uploaded sources in model training.

• AI technologies as a programming tool
  • For use in low-stakes business processes, for example to reduce the required time for one-off data or content manipulation tasks, using AI as a force multiplier in programming.

Version 1.0 25/08/25