The Intersection of AI and the Energy Sector: A Look into the Future

You’ve probably been at least vaguely aware of the media attention to the Chat GPT website which uses artificial intelligence (AI) to provide human like text responses to pretty much any question you might ask. We thought we’d take the opportunity to explore AI and the impacts it might have on the energy world. In fact, we asked Chat GPT a few questions and most of this article is based on chat GPT responses. The graphics come from Chat GPTs sister app DALL-E. We challenge you to tell the AI and human written text apart!

The integration of AI into the energy sector has the potential to influence the way we generate, distribute, and consume energy, possibly making these processes more efficient and reliable. AI is a combination of a variety of technologies, with a central part being machine learning, which is using complex algorithms applied to large datasets to identify patterns in data and draw conclusions, and AI can use this to give suggestions and make decisions.

This image was created with the assistance of DALL·E 2

Energy production

AI and machine learning could be used to optimise energy production and distribution by balancing input of multiple energy sources more effectively. This will become increasingly important as we become more reliant on renewables with variable output, such as solar and wind, allowing for a more consistent flow of energy into the grid. Additionally, AI could help recognise and predict possible faults by constantly analysing the variety of data sources, thus allowing for easier management of large-scale renewable production, as less direct management of systems will be needed.

AI Energy management systems

Savings in energy use will likely be possible with AI energy management systems for buildings, with heating, cooling, and lighting systems optimised with minimal human input needed. Patterns of usage could be automatically analysed for any inefficiencies or unexpected spikes in energy usage, and internal systems adjusted accordingly. This has the potential to drastically increase the energy efficiency of buildings by quickly identifying areas of concern, reducing usage and emissions in the process. This will save the time spent assessing energy usage, allowing for focus to move towards future of efficiency changes to the building, which could also be recommended by the AI energy management system. In addition AI could be managing on-site battery storage helping to control electric heating and EV charging systems, linked to on-site solar generation, to minimise electricity costs and respond to signals to reduce demand on the national grid.

This image was created with the assistance of DALL·E 2

Energy trading

Energy trading may also be impacted by AI, with huge amounts of historical trading data allowing for predictions of future changes in the market. Algorithms are used throughout the trading industry, and there are currently multiple AI-managed stock portfolios, and in the future, AI has the potential to operate will much less oversight. However, the application is extremely complex, with issues such as accountability for trades, as well as a lack of understanding of how AI trades will impact markets, which may limit AI application in this area in the future.

Limitations

One key limitation of the application of AI is that large datasets are required, and the energy sector needs to be fully digitised, so that more data can become available. Without this data, it is impossible to apply the functions of AI to the sector. The accuracy of this data is also key, as this is the basis for all decisions and information produced by AI systems. The accessibility of this digitisation is also important, as it must be affordable and make economic sense for businesses and homes to implement such changes. UK Government along with Ofgem and Innovate UK commissioned an Energy Data Taskforce which will report its recommendations to Government around setting modernisation and best practice standards for data in the energy industry.

Furthermore, general issues with AI such as accountability are also key factors to consider. If an AI makes a decision in an energy trade, and this ends in negative way, who is held responsible? It may also become difficult to understand the ‘thought’ process of the AI, as it will continue to learn on new information and data, and this fast development may accelerate beyond our understanding.

Overall, AI has the potential to save significant amounts of energy, and create more resilient energy grids, but it should be used cautiously, regulation must be fast evolving to keep pace, and a human element to decision-making should remain as a safeguard.

This blog was written by Freddie Williams, Zero Carbon Researcher.