Al and climate change – friends or foes?

In the rapidly evolving landscape of artificial intelligence and associated emerging technologies, the excitement about their transformative potential is palpable. From revolutionising health care with AI diagnostics to enhancing efficiency in manufacturing through automation, these technologies promise a future that is innovative and productive. However, as we enthusiastically chart this technological ascent, we must address a critical aspect that often remains underexplored: sustainability. The clean energy transition is imperative in our fight against climate change, yet the development and scaling up of new technologies, particularly AI, pose both challenges and opportunities. AI's significant energy consumption and its reliance on carbon-intensive materials for hardware construction cannot be overlooked. Recognising these challenges is only the first step – as we move forward, it is essential to develop strategies that not only mitigate these issues but also leverage AI to promote sustainability. One of the fundamental approaches to mitigating the environmental impact of AI is enhancing energy efficiency in data centres, where much of AI's heavy lifting occurs. Tech companies like Google and Facebook have already made strides by powering their data centres with renewable energy and employing AI itself to optimise energy use. Companies venturing into AI and other tech innovations can adopt similar strategies, such as using energy-efficient processors and advancing towards carbon-neutral data centres. AI can be instrumental in optimising energy usage across various sectors, from smart energy grids to dynamic routing in logistics, thereby reducing overall carbon footprints. Digital carbon emissions, stemming from our extensive use of technology, present a substantial environmental challenge. Every byte of data flowing through data centres and every device – from smartphones to edge computing sensors – require energy that is often derived from non-renewable sources. The enormity of these emissions can be traced back to the global reliance on digital technologies that are not yet fully sustainable. Addressing this issue demands a concerted effort at the enterprise level, starting by ensuring that data centres operate on renewable energy. However, transitioning to renewable power involves overcoming significant hurdles such as the need for specialised personnel, the procurement of advanced equipment, and financial investments that can stretch into billions and take years to actualise. Although these supply-side constraints are daunting, with thoughtful planning and long-term investments they not insurmountable. However, given the urgency of fighting climate change, faster solutions are also needed. An alternative strategy focuses on the demand side, specifically through the adoption of low-carbon coding or green coding practices. This approach aims to reduce the computational power required for each line of code, thereby decreasing the energy demands on computer and graphics processing units. Techniques such as regular maintenance to eliminate obsolete or excessive code, and rethinking code architecture, are pivotal. Although initially resource-intensive, embedding green coding practices within the core operations of a company promises extensive, long-term benefits. Not only does it reduce immediate carbon footprints, but it also cultivates a sustainable corporate culture poised to lead through the next century. The surge in demand for AI technologies, particularly neural processing units and related semiconductor chips, is also driving an increased need for raw materials such as silicon and rare earth elements like lanthanum and cerium. This heightened demand is likely to accelerate mining activities, potentially leading to more environmental degradation in the short term. Furthermore, the rapid expansion of AI technologies could result in increased electronic waste, mirroring concerns like those associated with the eventual disposal of decommissioned solar panels. The parallel with solar panels is instructive, highlighting the broader issue of electronic waste generated by advanced and emerging technologies. Addressing these material needs responsibly requires a comprehensive strategy that includes enhancing recycling technologies to manage end-of-life electronic products and developing new materials that can minimise environmental impacts. Such efforts will be crucial in mitigating the ecological footprint of AI's necessary materials, ensuring that the technology's benefits do not come at the expense of environmental sustainability. The role of regulatory frameworks in guiding AI development towards sustainability cannot be understated. Governments and international bodies can play pivotal roles by setting standards and benchmarks for sustainability in technology. This includes regulations around energy use, carbon emissions and waste management in the tech industry. Additionally, fostering collaboration between the public sector, private entities and academic institutions can drive innovation in sustainable practices. Initiatives such as the AI for Good platform by the UN are excellent examples of how collaborative efforts can harness AI for environmental sustainability. For AI and emerging technologies to develop sustainably, understanding and engagement across all levels of society are necessary. This means not only equipping developers and companies with the knowledge and tools to create sustainable technologies but also educating the public about the benefits and potential environmental impacts of these technologies. Greater awareness can drive consumer preference for sustainable products, thereby encouraging companies to prioritise green strategies. Perhaps the most optimistic aspect of AI in the context of sustainability is its potential to solve complex environmental problems. From monitoring deforestation and biodiversity loss using AI-powered satellites to developing algorithms that improve renewable energy storage, AI can be a formidable tool in our environmental conservation arsenal. Thus, while it presents certain challenges, AI also holds key solutions to some of the most pressing environmental issues. In conclusion, while the rapid development of AI and emerging technologies brings certain environmental challenges, it also offers unprecedented opportunities to drive sustainable practices. By integrating energy efficiency, lifecycle management, regulatory co-operation and public engagement into the development of these technologies, we can ensure that our technological progress is both innovative and sustainable. As we stand on the brink of this technological revolution, let us ensure that this progress is in harmony with our environmental responsibilities, paving the way for a sustainable future for all.