Artificial Intelligence – the next frontier for local energy transition

From drones, self-driving cars and robot vacuum cleaners sweeping our homes to chatbots answering our service requests, artificial intelligence (AI) is already appearing more and more in our daily city lives. But the advent of AI won’t end there : as energy systems are becoming decentralised, decarbonised and digitalised, artificial intelligence is set to emerge as the next frontier for local energy transition. What will this mean for cities ?

Artificial intelligence, which is the ability of machines to perform at the level of human intelligence, has been around for decades. In 1997, IBM developed the supercomputer “Deep Blue”, which managed to beat the world’s best chess player (Garry Kasparov). It had been trained by chess experts for 3 years. At that time, humans had to give a helping hand to machines for them to succeed. Today, programs have evolved exponentially : using trial and error, they can learn by themselves and no longer require human interaction. The global AI race has accelerated in the last few years, with US and Chinese tech giants at the forefront, while Europe is still lagging behind.

According to researchers and energy industry analysts, some of the key dimensions of AI are expected to have a dramatic impact on our cities and their energy systems :

  • deep learning, i.e. machines learning on their own by spotting patterns in data sets, 
  • algorithms that improve the more data they analyse, 
  • reinforcement learning, i.e. machines making experience-driven decisions, and 
  • the “Internet of Things”, i.e. networking buildings, vehicles, appliances, etc.).

While the use of AI is still in its infancy with regards to local energy systems, many possible applications have already been identified. Here are a number of options on how AI could be used in the local energy transition :
Autonomous grid management : AI as the future “brain” of the smart grid
Accurate predictions of renewable energy production
Optimising energy usage and consumption (demand management)
Improving management and maintenance of energy infrastructure
Revolutionising urban transport systems

What opportunities can AI applications bring for local energy systems ?

AI has the potential to change the future of local energy ecosystems and to support cities in delivering on their energy transition ambitions. 
AI systems are able to continuously collect and process large amounts of data from millions of smart sensors, meters and energy production facilities. They can thereby allocate energy resources, manage supply and demand and ensure an efficient and cost-effective operation of the grid.

AI software can also increase the uptake of renewables by addressing their intermittency through providing more accurate forecasts of wind and solar production. By combining data from wind and solar farms, weather stations, pollution monitors and local satellite reports, AI can identify patterns within these data sets and make accurate predictions that reduce the need for capacity mechanisms powered by dirty fossil fuels.

To succeed in their energy transition, cities not only need more renewables, but must first and foremost become more energy efficient.AI systems can, for example, monitor and learn from the energy consumption behaviour of individuals and businesses to improve their energy efficiency. Smart thermostats automatically adjust the temperatures of homes according to their occupants’ habits and bring significant monthly energy savings as a result.

Moreover, AI has the potential to improve the management and maintenance of energy infrastructure by constantly observing equipment and detecting failures before they happen, thereby saving costs and preventing blackouts. Furthermore, AI platforms can also be used to improve operation and cut costs in the water sector, as a pioneering project by the Australian city of Melbourne shows.

Finally, AI could optimise and revolutionise the way urban transport systems operate. AI-driven traffic systems could significantly reduce congestion, dynamically adjusting the signal timing of traffic lights as they observe and learn from traffic flows. As a further step, these intelligent traffic systems could communicate with self-driving cars to increase road safety and reduce the number of accidents.

Potential threats of using AI in the local energy transition

The advantages of applying AI in local energy are manifold, but its use is not without risk. Self-driving cars tested in the US have already caused several deadly accidents, casting doubt on their reliability. Moreover, AI could be maliciously used by those that program it to attack local energy infrastructures or cause blackouts, for example. Using AI also raises ethical questions : how can we trust that those who program AI solutions have our best interests at heart ? Who has access to the data and controls of these AIs ? And who is accountable when an AI – a machine – makes important decisions on its own that could threaten the security of our local energy systems and cities ?

Before its widespread deployment, it will be key to properly regulate and design ethical rules for using AI in cities’ energy transitions to ensure its safe and effective application. The first steps are being taken in that regard in Europe, as the European Commission will propose ethical guidelines for AI in the EU by the end of year. Energy Cities will closely follow the developments of AI in the energy sector and further analyse how it can support the local energy transition of cities.

Further reading



Publication date

August 7, 2018