Pafos Press 8 February 2026
In recent years, artificial intelligence (AI) has dynamically invaded our daily lives, radically transforming the way we communicated and thought until today. As part of the “Young Journalists for the Environment” program, a survey was conducted with questionnaires among students, parents and teachers aged between 15 and 75 in the Geroskipou area of Paphos. As can be seen from the results of our survey, a high percentage of respondents (73%) while using artificial intelligence very often in their daily lives (image 1), are not aware that it is related to the climate crisis (image 2). The results of our survey served as an incentive to inform and raise awareness of the local community on this issue. Since behind this “smart” technology, there is an environmental footprint related to the climate crisis that is rarely discussed and is worth knowing (image 3).
The development and use of Artificial Intelligence has significant environmental impacts through energy consumption and greenhouse gas emissions. In particular, training an AI model, i.e. feeding it with large, high-quality data sets so that it can learn patterns and make predictions over time, and running large AI models requires enormous computing power, which in turn consumes a lot of electricity and increases CO₂ emissions, especially when this energy is generated from fossil fuels [1]. Research has shown that training a single large language model can consume as much energy as five cars over their entire lifetime, including their manufacture and use [8].
Every time we use an AI application, from a chatbot to an image creation tool, we are activating powerful servers that consume energy 24 hours a day. While the user experience seems “intangible”, the computational processes behind these services leave a significant environmental footprint. In fact, many times this energy comes from non-renewable sources, burdening the environment with CO₂. The continuous operation of such systems, such as online AI services, adds new amounts of energy to the already burdened energy footprint of the digital world every day [8].
At the same time, in addition to the energy consumption for operating and training AI models, the management of the huge data sets that are stored and often unnecessary, leads to the increasing need for larger data centers that consume even more energy. Furthermore, uncontrolled data production exacerbates the problem, as the increase in raw information collected by businesses in combination with AI models leads to the consumption of more and more energy, resulting in the exacerbation of the greenhouse effect and the climate crisis [2].
One of the key approaches to address the environmental impacts of the use of Artificial Intelligence is the green digital transformation. According to UNESCO, the enormous energy requirements associated with the training and use of large AI models require a change in the way artificial intelligence infrastructures are designed and used. Specific measures such as reducing the size of models with compression techniques that reduce the size of data by saving space and transmission time, contribute to the elimination of redundant information and require much less energy [4].
An additional solution to the problem is the design of more efficient and “lightweight” AI models, the appropriate model must be selected for each task, so that small specialized models replace large Large Language Model (LLM) models that use various deep learning techniques and big data to produce text in a way that resembles human speech. At the same time, reducing the complexity of the commands we give to various applications – tools of artificial intelligence, can reduce energy consumption by up to 50% [7].
The Organization for Economic Cooperation and Development (OECD), in a special report on the environmental footprint of Artificial Intelligence, emphasizes that it is necessary for governments and international organizations to create systematic methods of measurement and transparency for the energy and other resources consumed by AI systems. They propose the establishment of standards for measuring the energy and environmental footprint of AI. They also propose expanding data collection for the accurate assessment of greenhouse gas emissions, not only at the level of operation but also at the level of production and disposal of the equipment. Finally, they propose transparency and sharing of measurements so that companies, countries and communities in general can adopt best practices and shape policies with real data. In this way, artificial intelligence will become part of the solution to the climate crisis, not part of the problem [5].
The United Nations Environment Programme (UNEP) supports a set of practical policy and technical measures to reduce the negative impact of AI. In particular, the programme supports the adoption of renewable energy sources by data centres, the buildings that host huge volumes of computers, servers and storage/telecommunications infrastructure, which are the heart of the “cloud” and the digital economy, in order to immediately reduce greenhouse gas emissions [3]. At the same time, it supports the mandatory disclosure of the environmental footprint of artificial intelligence products so that users, investors and regulators can compare and choose more environmentally friendly solutions [6]. We as citizens should not reject AI but when we use it, we should develop more responsible and sustainable practices such as: digital frugality, i.e. using AI services only when really needed, choosing AI services that operate on renewable energy sources where possible.
In conclusion, Artificial Intelligence, although it offers significant potential, can lead to serious environmental challenges. However, the negative impacts can be addressed with targeted practices. Informing and raising public awareness about the environmental consequences of the use of artificial intelligence will play a catalytic role in addressing the climate crisis. With proper use, artificial intelligence can evolve into an environmentally responsible tool that not only reduces its own environmental footprint, but also contributes positively to addressing the climate crisis.
Bibliography:
1. Navigating the environmental impact of AI
2. Everyone must understand the environmental costs of AI
3. How AI Policy can Accelerate Climate Action
4. Green Digital Transformation
5. Recommendation of the Council on Digital Technologies and the Environment
6. Artificial intelligence: How much energy does AI use?
7. Artificial intelligence and energy consumption: How AI systems affect data centers worldwide
8. The Environmental Footprint of Artificial Intelligence (AI)
Figure 1: Results of a survey with the percentages of people who use artificial intelligence in their daily lives.
Figure 2: Results of a survey with the percentages of people who believe that artificial intelligence is related to the climate crisis.
Figure 3: Title of the project: “The ala carte “taste” of an AI menu!” (Artistics, Christos Potsidis, Professor of Art).
Student group:
Theophanous Myrianthi B31
Kaiser Herodotus B31
Neofytou Georgia B32
Professor in charge: Melpo Tryphonos
