Posted in | News | Energy

Assessing the Strengths and Limitations of LLMs in the Energy Sector

According to new research carried out by scientists at Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS), large-language models (LLMs) could play a significant role when it comes to co-managing certain aspects of the grid, such as emergency and outage response, crew assignments, and wildfire preparedness and prevention.

Assessing the Strengths and Limitations of LLMs in the Energy Sector

Image Credit: Jamie Jin/

The benefits and drawbacks of large-language models in industries such as manufacturing, healthcare, and education have been extensively addressed. However, the same cannot be said about energy. This has thus raised questions as to whether ChatGPT and other LLMs could be used to operate and maintain the energy grid.

However, before LLMs are used in the field, we must overcome some of the security and safety issues that are limiting its widespread adoption.

There is so much hype with large-language models, it’s important for us to ask what LLMs can do well and, perhaps more importantly, what they can’t do well, at least not yet, in the power sector. The best way to describe the potential of LLMs in this sector is as a co-pilot. It’s not a pilot yet — but it can provide advice, a second opinion, and very timely responses with very few training data samples, which is really beneficial to human decision making.

Le Xie, Study Corresponding Author and Professor, Department of Electrical & Computer Engineering, Texas A&M University

Using GPT models, the study team—which included engineers from grid operator Midcontinent Independent System Operator and Houston-based energy supplier CenterPoint Energy—explored the potential of LLMs in the energy sector and found both advantages and disadvantages.

The strengths of LLMs, such as their capacity to learn from sparse data, assign tasks to embedded tools, generate logical responses to prompts, and work with non-text data like pictures, could be used to carry out tasks like identifying broken equipment, forecasting electricity loads in real-time, and evaluating wildfire patterns for risk assessments.

However, there are several obstacles to overcome before applying LLMs in the energy industry. One significant challenge is the lack of grid-specific data for model training. Critical information about the US power grid cannot be made public and is not available to the general public due to security concerns. Additionally, the absence of safety guardrails presents another problem.

The power grid, similar to autonomous vehicles, must prioritize safety and maintain a significant safety buffer while making real-time decisions. Li also believes that LLMs should enhance their ability to deliver reliable solutions and communicate their uncertainty clearly.

We want foundational LLMs to be able to say ‘I don’t know’ or ‘I only have 50% certainty about this response’, rather than give us an answer that might be wrong. We need to be able to count on these models to provide us with reliable solutions that meet specified standards for safety and resiliency.

Na Li, Winokur Family Professor, Electrical Engineering and Applied Mathematics, John A. Paulson School of Engineering and Applied Sciences, Harvard University

All of these problems provide engineers with a path for future development.

As engineers, we want to highlight these limitations because we want to see how we can improve them. Power system engineers can help improve security and safety guarantees by either fine tuning the foundational LLM or developing our own foundational model for the power systems. One exciting part of this research is that it is a snapshot in time. Next year or even sooner, we can go back and revisit all these challenges and see if there has been any improvement.

Na Li, Winokur Family Professor, Electrical Engineering and Applied Mathematics, John A. Paulson School of Engineering and Applied Sciences, Harvard University

Subir Majumder, Lin Dong, Fatemeh Doudi, Yuting Cai, Chao Tian, Dileep Kalathil of Texas A&M University; Kevin Ding of CenterPoint Energy; and Anupam A. Thatte of Midcontinent Independent System Operator are the study co-authors.

Journal Reference:

Majumder, S., et. al. (2024) Exploring the capabilities and limitations of large language models in the electric energy sector. Joule. doi:10.1016/j.joule.2024.05.009

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.