Will AI Make or Break the Energy Transition? Insights from Nature and Climate House

At Empower, we work with purpose-led organisations and nonprofits to harness digital tools for positive social and environmental impact.

As AI increasingly shapes the digital landscape, understanding its role in addressing global challenges like climate change is crucial for our clients who are working on the frontlines of environmental and social change. 

The evolving applications of AI in early warning systems, resource optimisation, and breakthrough research directly align with the kind of transformative work our partners are doing, making these insights invaluable for both our strategic approach and the causes we support.

This is why we were keen to attend the recent “Will AI Make or Break the Energy Transition?” event at Nature and Climate House, which brought together experts to discuss whether AI will be a critical enabler for achieving net-zero or if its own energy demands pose a threat to energy security.

The panel featured insightful perspectives from:

• Dan Travers, Co-Founder of Open Climate Fix
• Leila Toplic, Chief Communications & Trust Officer at Carbon Future
• Sims Witherspoon, Climate Action Lead at Google DeepMind

A fascinating conversation unfolded and the discussions highlighted several key themes around the energy transition diving deep into the role of AI in shaping our energy future. 

AI’s potential for positive impact

Leila Toplic highlighted several critical areas where AI is already demonstrating significant positive impact. 

One of the most pressing applications she addressed was the use of AI in early warning systems. She noted the United Nations’ statistics indicating that AI-powered early warning systems could reduce the impact of disasters by up to 30% when effectively implemented. 

To illustrate this, she shared that AI flood alerts are now accessible across a hundred countries, reaching 700 million people. This provides timely information that is essential for preventing loss of life, protecting livelihoods, and safeguarding infrastructure.

Toplic also emphasised the critical intersection of climate and health, citing a specific example from Rwanda. In this instance, an AI-based system is being used to predict malaria outbreaks by analysing environmental factors, such as glyphosate volatility and its impact on mosquito patterns. 

This predictive capability enables the strategic prepositioning of resources, allowing for more effective prevention and management of malaria outbreaks.

These examples serve to demonstrate how timely, AI-driven information can empower people to take action, significantly reducing negative impacts and improving outcomes.

Frameworks for applying AI

Sims Witherspoon introduced a valuable three-part framework to help understand how AI can be effectively applied to various challenges, particularly in the context of the energy transition. 

1. Understand

AI excels at processing vast amounts of data(far more than any human could) and extracting patterns that might otherwise be missed. This capability allows for a much deeper understanding of complex problems and reveals insights that can inform better solutions. Early warning systems and forecasting were cited as prime examples of this. 

2. Optimise

Witherspoon expressed excitement about optimisation problems, as AI is particularly well-suited to tackling them. In the context of the energy transition, optimising existing systems and infrastructure is critical. For example, the national grid, which is vital for current electricity supply, cannot simply be replaced overnight. AI can provide software-only solutions to optimise these existing systems while more sustainable technologies are developed in parallel. 

3. Accelerate

She also emphasised that AI can accelerate breakthroughs in science and the path to solutions. She illustrated this with an example from fusion energy research, where DeepMind has developed a reinforcement learning system capable of controlling plasma in a real-world tokamak fusion reactor. This is significant because controlling plasma is a critical hurdle in achieving nearly inexhaustible carbon-free energy from fusion.

Regardless of the industry, be it energy or otherwise, this framework of “Understand,” “Optimize,” and “Accelerate” can guide efforts to identify where AI can be most effectively used.

By asking what problems need understanding, optimiwation, or speed, organisations can discover high-probability areas for AI application.

AI challenges and considerations

Nonetheless, Dan Travers noted several challenges and considerations that must be taken into account when implementing AI solutions. 

One significant concern raised was the potential for AI itself to consume substantial amounts of energy. It was acknowledged that finding a balance between the benefits of AI and its energy footprint is essential. 

Another critical point was the importance of data availability and quality. AI relies on data, and questions about the accessibility, quality, and privacy of energy data must be addressed. It was emphasiwed that well-regulated data sharing is crucial for effective AI applications. 

Ensuring equitable access to AI benefits, particularly in low and middle-income countries (LMICs), was highlighted as a key consideration. Educational and training programs are needed to bridge the skilling gap and prevent AI from exacerbating existing inequalities. 

Climate misinformation

The panel also touched on the issue of misinformation, specifically AI-generated climate misinformation, and the necessity of using credible data to train AI models. 

A fundamental question was raised about which problems should be prioritised for AI solutions: should AI focus on critical issues like early warning systems and health risks, or will it primarily be used for less essential applications like advertising? 

The energy transition requires holding two competing truths at the same time and recognising that choices are not always binary.

There’s an ongoing need to understand the relationship between AI and human labor: will it help us perform our jobs better, or will it replace jobs? 

AI and the Energy Transition: key takeaways

• AI has immense potential to drive positive change in the energy transition, from improving efficiency to accelerating scientific breakthroughs.
• It’s crucial to address the challenges related to AI’s energy consumption, data management, and equitable access.
• Defining clear objectives and benchmarks is essential for successful AI implementation.
• Partnerships between AI practitioners and domain experts are vital for solving complex problems.

The event highlighted that the role of AI in the energy transition is complex and multifaceted. It requires careful consideration of both its potential benefits and potential risks. As the discussions concluded, it became clear that this is an ongoing conversation with many more questions to explore as we navigate this evolving landscape.