Automated Energy Futures: Between Opportunity and Risk

Screens flicker, as data rolls and recalibrates. Artificial intelligence (AI) pulls up a snapshot of data facilities from thousands of sensors. This data is fed into ‘deep neural networks’, which are basically computational processes using sophisticated mathematical modelling to make predictions. In this case, the predictions concern how different choices will affect future energy consumption. The city is London. The setting is DeepMind, the pioneering British technology company. The context is global, with DeepMind deploying its cloud-based AI for client Google to directly control cooling in the data centers that power services including Google Search, Gmail and YouTube. The results appear remarkably effective. Google has in recent years been able to directly harness DeepMind’s AI to reduce its energy consumption by roughly 30%. Google’s deployment of DeepMind’s AI for advancing energy efficiency would not be possible if not for a team of data operators and AI technicians meticulously calculating the best course of potential actions to cut energy use. But since 2018, advancements in automation have made it possible for AI to directly control cooling in Google’s data centers without human input. Such AI-engineered energy efficiency is a powerful vision of smart infrastructures for the future of energy security, administered and managed daily through automated intelligent machines. This example drawn from the emergent uses of AI for energy security is part of a new agenda for politics and public life today, an agenda that concerns two interconnected domains of transformation. On the one hand there is the extensional spread of AI, which increasingly plays a key role in realizing a low-carbon (and possibly post-carbon) climate and energy future. On the other hand, but immediately bound up with the first, are processes of advanced machine learning where algorithmic technologies fuel the erosion of energy security and, so some argue, contribute to the burning of the planet. We are currently living in a world that, as far as advances in AI are concerned, looks split in terms of such opportunities and risks. Most of those who have studied advances in AI have emphasized the resulting transformed connections between climate change and energy security. For instance, Anthony Giddens argues that “the realm of technology is the most important domain where the theorem applies that the quantum leap in power that has created dangers for us can allow us to meet them”[1]. Meeting the immense challenges of climate change through technological innovation such as AI, says Giddens, must involve both concrete and radical measures. This emergent discourse of AI-powered low-carbon climate futures presupposes that new technologies will somehow remedy the problem of global warming through providing novel ways of generating alternative energy supplies. From reducing greenhouse gas emissions and especially CO2 emissions to reversing the world’s rising sea levels, technological advances in AI and connected digitalization will be vital to our chances of confronting global climate change challenges. The impact of technology in promoting low-carbon lifestyles is likely to be bound up directly with AI breakthroughs, machine learning experiments and new algorithmic-driven social and environmental practices. Some authors have argued that AI is increasingly becoming the central driving force of energy adaptation throughout the economy as a whole[2]. What is being underscored here is that AI has been marshalled to advance energy efficiency in a variety of industrial settings, particularly with the spread of smart cities, but also within the energy sector itself. Google’s utilization of DeepMind’s AI is one key example of this trend. But other critics have noted that trends alone do not equal transformations. From this standpoint, AI can only contribute to the realization of a new post-carbon energy system if it is systematically implemented around the world. Such epochal change would demand the rolling out of radically novel technological solutions to create more ecologically sustainable societies, along with the implementation of new sources of energy. But, again, many critics view all this as simply fanciful. In all of this, possibility and actuality jostle together uneasily. The techno-optimists argue that new technologies will offset environmental damage and foster low-carbon or post-carbon energy solutions around the world. Some even argue that advances in machine learning algorithms are likely to be radically transformative, impacting almost all fields of social and economic life in the future. But such projected breakthroughs in AI are, by definition, the least predictable and almost impossible to assess in terms of policy considerations and socio-economic development. On the other hand, the sceptics highlight that AI technologies are already bound up in destructive ways with energy sources of carbon capitalism. From this angle, it is difficult to see how these technologies might ever transfigure or shift beyond carbon energy-intensive practices. Yet it is, again, crucial that we bear in mind the open-endedness of social futures. If AI research and innovation can contribute to new ways of storing electricity, it follows that this would have an enormous impact on energy security. Various developments concerning electricity storage such as flow batteries, supercapacitors and power conversion systems are key as possible storage technologies and, when combined with advances in AI, offer possibly progressive ways forward towards a pervasive powering down to low-carbon economies and societies. [1]Anthony Giddens, The Politics of Climate Change, Cambridge: Polity Press, 2009, p.230. [2]For further discussion see Anthony Elliott, Making Sense of AI: Our Algorithmic Worlds, Cambridge: Polity, 2022.

{{ currentMenugamenu.label }}

{{ currentSubMenugamenu.label }}

{{ card.title }}

{{ currentSubMenugamenu.numbers.title }}

Automated Energy Futures: Between Opportunity and Risk

Automated Energy Futures: Between Opportunity and Risk

Anthony Elliott

Robotics and automation: opportunities and risks

Senior researcher at the Italian Institute of Technology in Genoa

Mechanical and biomedical engineer