- Balancing the benefits of technologies with their environmental impact or addressing the paradox that while technology offers innovative solutions to environmental challenges, it also contributes to e-waste, global warming, and energy consumption. It remains challenging to reconcile both aspects, harnessing the technological potential for positive change while minimising its negative consequences.
- Democratisation in access to and use of technology has also brought an overuse that is not always fully justified. For example, on average, a ChatGPT query requires nearly 10 times as much electricity to process as a Google search. This challenge requires educating people on the responsible use of AI regarding energy efficiency and its use according to their needs.
- Sustainability and Circular Economy: We must move from the “take-make-dispose” paradigm to a more sustainable approach based on a circular economy, where digitalisation can play a key role.
- Increase ICT sustainability at large: This challenge entails a set of actions, like raw material optimisation (develop and promote the use of sustainable materials, implement circular economy principles, including recycling and reusing materials to minimise the extraction of new raw materials), reduction of wastage (extend the lifespan of ICT devices and reduce e-waste), energy consumption reduction (energy-efficient technologies and practices across all stages of the ICT lifecycle, increase the adoption of renewable energy sources in powering ICT operations), water consumption reduction (closed-loop water systems to reduce the need for freshwater), policy and regulation, collaboration and consumer awareness and behaviour.
- Energy use requirements for data-processing technologies: As data processing demand grows exponentially, driven by advancements in AI and big data analytics, the energy consumption of data centres and related ICT infrastructures has become a significant concern. The rapid expansion of these technologies has led to a ballooning in energy use, contradicting global efforts to reduce carbon footprints and combat climate change (an average hyperscale data centre consumes between 20 to 50 MW per year, like the amount of energy theoretically sufficient to power up 37,000 homes, data centres represent 1% of the global energy demand and 0.3% of the global carbon emissions).
- Environmental impact of energy-intensive AI and large models training and interaction: while the rise of AI can help to address global warming and climate change, AI is also a great contributor to global warming. Large-scale AI systems, due to their energy-demanding model training and operations, leave a significant carbon footprint (the carbon footprint of training a single AI is equivalent to five times the lifetime emissions of an average car).
- Energy efficiency in HPC: Despite the efforts of EuroHPC (see dedicated call and green EuroHPC supercomputers in the previous section), reducing the impact of these computers remains a challenge.
- Sectorial environmental challenges: e.g., sustainable urban planning and management, manufacturing (low-impact manufacturing processes and reducing the carbon footprint of production, Promoting supply chain transparency and accountability), etc.