Call for input AI Technical Sandbox
The IET has responded to the AI Technical Sandbox consultation for Ofgem.
AI is reshaping how industries and technologies operate, offering powerful tools for resilience, scenario modelling, and data-driven decision-making. However, it must be applied responsibly, with attention to risks and system vulnerabilities.
A sandbox could help demonstrate how AI can be used successfully and encourage other sectors to adopt AI by supporting safe experimentation, particularly in areas like dynamic pricing and energy management. It is important, however, for the sandbox to have access to representative energy system data to ensure accurate modelling.
This could be achieved by ensuring compatibility with existing digital twins within the national energy infrastructure. Digital twins enable testing of AI behaviours and help identify both beneficial and harmful outcomes, as well as real-world impacts. Ensuring high-quality data and ongoing validation is critical for accurate modelling and effective integration with AI systems.
The IET’s recommendations:
- Energy system data modelling – The sandbox needs to be designed to run projects more efficiently using up-to-date information throughout the process, leading to improved productivity.
- Raising awareness of carbon efficient AI models – As AI models have very large compute and storage demands, there will be a need for new data centres. The IET recommends a bronze, silver, gold standard for the approval of new data centres in the UK, based on a sustainability rating.
- Regulation – AI safety and the assessment of risk must go beyond the physical, to look at financial, societal, reputational and risks to mental health, amongst other harms.
- Data Governance – There should be firm rules on which data can and cannot be used to train AI systems. The AI Technical Sandbox is a good way to establish high quality and use of data.