The Sustainability Imperative for Responsible AI Innovation

lorenaflorian0
Oct 2
3 min read

Artificial Intelligence (AI) is rapidly transforming industries, revolutionising how we work, make decisions, and deliver value. From predictive analytics to generative tools, AI is now a cornerstone of digital transformation strategies. But as organisations race to harness its potential, a critical dimension is often overlooked; the environmental cost of AI.

At the recent AIPM 2025 Conference ‘Future in Focus’, while much of the conversation focused on AI’s role in enhancing project efficiency and data-driven decision-making, the environmental implications of AI remained largely absent.

The Environmental Cost of AI

AI systems, particularly large-scale models, are resource-intensive by design. Their environmental impact spans several key dimensions:

Energy Consumption

Training advanced AI models requires immense computational power. The energy required to train models like GPT-3 can exceed 1,200 megawatt-hours; comparable to the yearly electricity usage of over 120 American households. Even routine AI tasks, such as generating images or responding to prompts, consume significantly more energy than traditional computing tasks.

Greenhouse Gas Emissions

The energy powering AI systems often comes from non-renewable sources, contributing to greenhouse gas (GHG) emissions. As AI adoption grows, so too does its carbon footprint. The broader information and communications technology (ICT) sector, which includes AI, already accounts for at least 1.7% of global emissions—a figure expected to rise sharply as AI usage scales.

Water Usage

Data centres, which power AI infrastructure, require extensive water for cooling. With AI’s projected water withdrawal expected to reach 6.6 billion cubic metres annually by 2027, this demand could exacerbate water scarcity in vulnerable regions.

Hardware Lifecycle

AI’s physical infrastructure—servers, GPUs, and specialised chips—requires significant raw materials and energy to produce. The lifecycle of this hardware, from manufacturing to disposal, contributes to electronic waste and environmental degradation. Moreover, the construction and maintenance of data centres themselves can account for up to two-thirds of their lifetime emissions.

Why Green AI Matters

The good news is that sustainability and AI innovation are not mutually exclusive. By adopting mindful practices, organisations can reduce AI’s environmental impact while still reaping its benefits.

The benefits of integrating sustainability into AI strategies include:

Environmental protection: Reducing emissions and conserving resources
Economic efficiency: Lowering energy and cooling costs
Ethical responsibility: Aligning with values of environmental stewardship
Regulatory readiness: Preparing for evolving environmental compliance standards
Reputation and competitiveness: Meeting stakeholder expectations for sustainability

Practical Steps for Greener AI

Embrace Green AI Principles

Green AI refers to the development and deployment of AI technologies that minimise environmental impact through energy efficiency, reduced carbon emissions, and sustainable resource use. It aligns with the UN’s Sustainable Development Goals (SDGs) and offers a pathway to responsible innovation. This includes

Using smaller, more efficient models where possible
Prioritising algorithmic optimisation to reduce computational load
Leveraging cloud and edge computing to minimise energy use

Optimise Data Centre Efficiency

Investing in energy-efficient infrastructure and sourcing renewable energy for data centres can significantly cut emissions. Organisations should also consider the geographic location of data centres, favouring regions with cooler climates and access to sustainable energy.

Track and Report AI-Related Emissions

Integrate AI-specific metrics into sustainability reporting. This includes tracking energy and water usage, hardware lifecycle impacts, and emissions associated with AI operations. Transparency in these areas can drive accountability and improvement.

Incorporate Sustainability into AI Governance

Responsible AI frameworks should explicitly include environmental considerations. This means embedding sustainability into procurement, model development, and deployment decisions—ensuring that ethical AI is also environmentally responsible AI.

Educate and Empower Teams

Sustainability should be a shared responsibility across technical and non-technical teams. Providing training on the environmental impacts of AI and encouraging cross-functional collaboration can foster a culture of conscious innovation.

AI holds immense promise, but its benefits must not come at the expense of the planet. As stewards of digital transformation, organisations have a responsibility to ensure that their AI strategies are not only effective but also sustainable. By acknowledging the environmental footprint of AI and taking proactive steps to reduce it, we can build a future where technological progress and ecological stewardship go hand in hand.

References:

Litvinets, V. & EY Global. (2024, November 14). AI and Sustainability: Opportunities, Challenges, and Impact. EY. https://www.ey.com/en_nl/insights/climate-change-sustainability-services/ai-and-sustainability-opportunities-challenges-and-impact

Bartczak, J., & Block, S. (2025, June 1). How AI use impacts the environment and what you can do about it. World Economic Forum. https://www.weforum.org/stories/2025/06/how-ai-use-impacts-the-environment/

Bolón-Canedo, V., Morán-Fernández, L., Cancela, B., & Alonso-Betanzos, A. (2024). A review of green artificial intelligence: Towards a more sustainable future. Neurocomputing, 599, 128096. https://doi.org/10.1016/j.neucom.2024.128096