The emergence of Artificial intelligence is undoubtedly affecting sectors like that of global productivity, research, but also environmental outcomes. By now, we all know that it is a powerful tool… but to what extent? Can it actually be useful to do better?
AI is rapidly spreading throughout civilization, where it has the promise of doing everything from enabling self driving cars to navigate the streets to making more accurate hurricane forecasts. Artificial intelligence (AI) enable computers to reason on inputs and perform a variety of advanced functions, including the ability to see, understand and translate spoken and written language, analyze data, make recommendations, and more.
This technology operates through complex algorithms that process vast amounts of data to identify patterns, make predictions, and automate decision-making. At its core, AI relies on machine learning models, particularly deep learning, which uses artificial neural networks inspired by the human brain. These models require extensive training, often performed on specialized hardware such as Graphics Processing Units (GPUs) or Tensor Processing Units (TPUs), which accelerate computations through highly resource-intensive processes. Training large-scale AI models demands, in fact, significant computational power, leading to high energy consumption. Data centers that support AI operations require substantial electricity, often relying on non-renewable energy sources, and contribute to carbon emissions. They further rely on water both during construction and, once operational, to cool electrical components. Additionally, AI systems necessitate vast datasets, which involve storage infrastructure that further increases energy use and produce electronic waste, mostly mercury and lead.
According to the Energy Agency, a request made through ChatGPT, an AI-based virtual assistant, consumes 10 times the electricity of a Google Search. This comparison highlights the immense energy demands of AI systems. As more individuals and businesses turn to AI-driven tools for automation, decision-making, and problem-solving, the cumulative energy consumption will escalate rapidly. If left unchecked, this trend could lead to an unsustainable surge in electricity demand, exacerbating carbon emissions and straining global energy supplies. Without responsible development and regulation, AI’s growing dependency on high-performance computing has the potential to become an environmental threat. However, the technology also has the potential to tackle some of the world’s biggest environmental emergencies.
A comprehensive study published in Nature Communications analyzed the impact of AI on the 2030 Agenda for Sustainable Development. Even though there is still no published study assessing to what point AI might impact in achieving the goals and targets agreed, the findings indicate that AI could support 134 targets across all 17 SDGs, meaning 79% of them all.
It is still crucial to highlight that the other 59 may experience a negative impact from the development of AI. The technology has an impact on each one of the three pillars of sustainable development, Society, Economy and the Environment.
Within the Society group, the number of targets that could benefit from the help of artificial intelligence is around 80%: for goals like SDG4 on quality of education or SDG11 on sustainable cities AI could act as an enabler by supporting, for example, the creation of circular economies and smart cities that fully maximise their resources. Only the 38% of targets in the Society group can be impacted negatively, but we still need to keep them in mind: many of these are related to the difficulties that would be faced in the process of integrating technological improvements in countries with different cultural values and traditions. To work, in fact, artificial intelligence needs huge computational resources, and to realise this it is sufficient to compare the quantity of energy used by cryptocurrency applications like Bitcoin, that globally consume the same as some countries’ entire electrical demand.
Green growth of ICT technology is therefore essential, especially if we consider that some estimates suggest that the total electricity amount for the sector could go up to 20% of the total electricity demand by 2030, from 1% today.
Another important drawback is that this tool is traditionally based on the needs of nations in which AI is being developed, but not the whole world is the same: matters of transparency, ethical scrutiny and democratic control are dealt with very differently among countries, and this could potentially enable feelings of hate or rage. No matter how advanced we might perceive artificial intelligence to be, what is clear is that the road to become more useful than dangerous is still complex.
The technological advantages provided by AI can have a positive impact also on targets related to the Economy group: 42 of them, specifically, could benefit from the further development of ICT, while negative impacts are reported in the 33% of SDGs. What is most important, for this sector, is that data analysis -a resource on which we rely heavily- needs to be equally available in all countries, otherwise the risk of increasing the economical gap may be strengthened by the newly introduced inequalities. SDG8 on decent work and economic growth and SDG9 on industry, innovation and infrastructures considerably depend upon how we decide to develop and use artificial intelligence. Although there are more points in favour than against in this sector, trade-offs cannot be ignored: AI could have a negative effect on social media usage, customising the information to show to users depending on their personal needs or ideas. This could harm politics and affect social cohesion, enhancing conflict and inequality.
The last group we consider has to do with Sustainability, where AI could work to the benefit of 25 targets (93%). These benefits are deeply connected with the possibilities to analyze large-scale databases to develop plans and strategies to preserve the environment, not to mention the fact that the tool itself could support the understanding of climate change and provide models of its impacts and potential solutions. Objective-oriented techniques can be implemented to identify environmental trends taking place around the Earth, such as desertification or the rising of the sea level. On the other side, however, AI could also be conjectured to exploit resources and enlarge already existing gaps. The risks are present, of course, but up to this date it is important to underline that there is no strong evidence on misuses of information.
“The role AI can play in tackling environmental challenges, can be on a large scale – such as satellite monitoring of global emissions, or a more granular scale – such as a smart house automatically turning off lights or heat after a certain time” , said David Jenses, coordinator of the United Nations Environment Program’s (UNEP’s) Digital Transformation sub-programme. The difficulty in the future will be to fully adapt and control this tool to reach in a faster and safer way the goals agreed in the 2030 Agenda on Sustainable Goals, without forgetting about the environmental costs and damages AI itself creates.
While AI has the potential to drive sustainability solutions, its environmental footprint must be carefully managed to ensure that its benefits outweigh its resource costs. AI, therefore, stands at a crossroads between being a powerful tool for sustainability and a potential environmental burden. Its ability to optimize resource management and enhance climate modeling makes it an ally in tackling global challenges. However, its rapid expansion and the growing energy demands of AI-powered technologies pose serious risks that cannot be ignored. The future of AI in sustainability will depend on how effectively we balance its advantages with its drawbacks by developing energy-efficient models, prioritizing green computing, and ensuring equitable access to AI-driven solutions. If guided responsibly, AI could become a transformative force in achieving the Sustainable Development Goals, helping create a more resilient and environmentally conscious world. However, without proper governance and innovation in sustainable AI infrastructure, its unchecked growth could deepen inequalities and exacerbate environmental pressures. The key challenge moving forward will be to harness AI’s potential while minimizing its costs, ensuring that it contributes to a technologically advanced and ecologically sustainable future.
Written by Erica Laudicina and Caterina Molinari
References:
- https://sdg-action.org/can-ai-help-us-achieve-the-sdgs/#:~:text=The%20answer%20is%20 yes%20%E2%80%93%20potentially,more%20sustainable%20future%20for%20all.
- https://www.nature.com/articles/s41467-019-14108-y
- https://www.unep.org/news-and-stories/story/how-artificial-intelligence-helping-tackle-envi ronmental-challenges
- How Does AI Work? | HowStuffWorks
- Artificial Intelligence (AI): What it is and why it matters | SAS India
- AI has an environmental problem. Here’s what the world can do about that.
