IB1 Research Guidance: Using Large Language Models

Ross Crear — Thu, 13 Nov 2025 11:20:10 +0000

At Icebreaker One, we have observed a significant increase in the use of Large Language Models (LLMs) across the sectors we work in. It’s important to recognise that while LLMs might be widely adopted, they often lack appropriate oversight, which is why we’ve developed guidance for both internal use and for our partners.

The responsible use of LLMs includes understanding their energy use and the net zero implications associated with them. Both LLM model creation and usage are known to use a significant amount of energy and so this guide includes considerations for reducing energy consumption in LLM usage.

Consistent with our commitment to transparency, we have built upon the excellent work of mySociety’s AI Framework to produce a practical guide for internal use. We are publishing it so partners can understand our approach, adapt it for their own contexts, and engage with us where collaboration would be helpful. The guide addresses how to use LLMs responsibly as research tools while maintaining rigorous oversight and expert review.

Written by Paul Johnston

Read our full Guide to Using LLM’s here

research – Icebreaker One

IB1 Research Guidance: Using Large Language Models