Eurelectric is restructuring its working scene to awaken the power sector to the risks and benefits of artificial intelligence in energy

Artificial intelligence (AI) is currently a two-horse race between China and the US, with the later having attracted almost 70% of the world’s external investment in the technology. In an attempt to catch-up, Europe recently published a strategy to scale up public and private investments and provide a framework for the development of a technology we are yet to know if we can fully master. Eurelectric is also restructuring its working scene to awaken our industry to the risks and benefits of AI in energy. “The good news about AI is that it produces more results with fewer resources. Hence if we are smarter and focused on win-win type results, AI could help efficiently distribute the world’s existing resources like food and energy,” said Sophia, the AI robot, when intervening at the United Nations after having been granted Saudi citizenship. She showed no sign of nerves when making her initial address to diplomats and her subsequent interviews prove the width and scope of artificial intelligence. The idea behind AI and machine learning (ML) is simple, but revolutionary: a machine is able to display adaptive intelligence, where its behaviour is not pre-determined and it can adapt to data inputs. ML is used to enable automation and can increase potential in almost every application that was previously reliant on predictive data. Energy is no exception.

A smart move for utilities

Efficiency gains in the power sector resulting from the uptake of AI are indisputable. In many cases, AI brings cost reductions throughout the value chain and ultimately results in lower bills for consumers. The clearest use of AI relates to dynamic maintenance, where data collected from various units through sensors enables operators to predict fault and equipment failures. Looking at years of operational data, Google’s algorithm Deepmind submits advice on changes to make to individual units. This has led to a 15% reduction in energy use in the company’s data centres. Google has also announced ongoing discussions with the UK’s National Grid to better forecast demand and reduce the country’s energy use by 10%. Another promising example is the use of ML models to forecast weather conditions at specific generation units. This helps to predict their real-time power generation and makes it easier to integrate renewables into the network. On the consumption side, retailers use ML to reveal more clearly patterns in consumers’ behaviour and unlock demand-side flexibility. Meanwhile, chatbots and devices such as Amazon’s Alexa enable increased interactions with customers and allow for the development of more personalised usage profiles. But while AI promises better forecasting and increased demand-side flexibility, a mere 23% of European utility executives see AI as a strategic priority. The reasons for such a relatively late adoption of AI by utility leaders are multiple. First, utilities struggle to attract data scientists who are key for understanding and implementing AI technology. Second, there are the potential risks, a faulty response from an algorithm could result in potential power outages. Beyond providing better predictions and helping automation, AI can be a useful tool to aid complex decision-making and help provide solutions to the provision of clean energy for all. There is still a long way to go, however, before this becomes reality. In an increasingly decentralised landscape with fierce competition coming from outside the industry, AI is one more area where it is imperative to break down silos and ensure that we all work together for the greater good. To tackle these challenges, Eurelectric is creating a discussion platform where utilities can exchange about AI in a safe environment and team-up with start-ups and tech giants. This platform will enable actors at all layers of the energy system to collaborate and allow participants to understand how to respond to the development of AI and direct long-range investment cycles while stressing a culture of fail fast, move on. This platform should be a forum to develop a unity of views and of actions for European electric utilities and harvest the benefits of AI, allowing the power sector to successfully find its way in the global race.

The European response

Beyond the power sector, in an environment where the earliest adopters of AI are likely to become the next world leaders, the sense of urgency for Europe to wake up is real. While the US and China were pioneers in adopting national AI strategies, the EU has taken more time to adopt a coordinated approach to the technology, only recently, in an attempt to tackle the investment gap, did the European Commission present its AI strategy, stressing the need for public and private funding to reach a target of €20 billion a year over the next decade. As we place more of our decisions in the hands of a technology we do not fully master, ethical questions will become top priority for the European agenda. Some ML algorithms are so complex that their owners cannot fully grasp the results obtained and leave human beings with no choice but to blindly trust the system while taking responsibility for its decisions (some of them as crucial as allowing a self-driving car to decide who to save between passengers and pedestrians). The Commission intends to provide a framework for trustworthy AI, where clear ethical principles can be found in the form of a checklist when developing and using AI systems. It is also setting up an AI Alliance, a forum aimed at discussing all aspects of AI development and its impacts. Europe has proven many times that it is able to provide the right responses to drive competitiveness for its core industries. Questions of ethics, privacy and security are already given the highest priority under Europe’s largest research programme Horizon 2020. A European AI strategy that stimulates industrial competitiveness and provides privacy and trustworthiness would in the long run likely position the EU at the front of the AI race.

Do you have a thoughtful response to the opinion expressed here? Do you have an opinion regarding an aspect of the global energy transition you would like to share with other FORESIGHT readers? If so, please send a short pitch of 200 words and a sentence explaining why you are the right person to deliver this opinion to opinion@foresightdk.com.