AI’s Sustainability Balancing Act: What’s at Stake?

As we stand at the intersection of technological advancement and environmental urgency, the role of artificial intelligence (AI) in advancing the sustainability agenda is becoming more pronounced than ever.

In my previous article, I explored how AI can drive sustainable innovation while addressing its inherent risks, such as energy consumption and misinformation. Today, I want to delve deeper into the dual nature of AI, its transformative potential to solve environmental challenges and its significant resource demands, and how we can strike a balance to ensure a sustainable future. 

To ensure this article captures the full spectrum of perspectives, I’ve interviewed global senior thought leaders from sustainability, AI, supply chain leadership, and circular economy fields. Their insights have shaped this discussion, offering a comprehensive view of the opportunities, challenges, and strategies for integrating AI into sustainability efforts. 

As AI technology continues to expand across industries and regions, its potential to drive environmental sustainability is becoming increasingly evident. From optimizing energy use to enabling circular economies, AI is already proving its worth as a tool for environmental stewardship.

Initiatives like the AI for Good movement, led by organizations such as the United Nations, showcase how AI can be applied to support the Sustainable Development Goals (SDGs). Many of these goals, including Goal 13: Climate Action, are directly linked to protecting the planet and addressing environmental challenges. The infographic below illustrates how AI for Good initiatives specifically align with the UN’s SDGs, highlighting key environmental objectives such as optimizing renewable energy, enhancing climate monitoring, and promoting responsible resource management.

The European Parliament’s Think Tank estimates that AI could reduce global greenhouse gas (GHG) emissions by 1.5-4% by 2030, directly supporting the United Nations’ Sustainable Development Goal (SDG) 13 on climate action [1]. For instance, Google DeepMind reduced data center cooling energy use by 40%, showcasing AI’s potential to drive systemic efficiencies [2]. 

In industries like consumer goods, AI is revolutionizing sustainability efforts. In cosmetics, AI predicts the environmental impact of ingredients, enabling brands like L'Oréal to create sustainable beauty products [3]. In food and beverage, companies like Nestlé are using AI to optimize supply chains, reduce food waste, and develop plant-based alternatives with lower carbon footprints [4]. And in pharmaceuticals, AI is accelerating drug discovery while minimizing lab waste and energy use, as seen in Pfizer ’s use of machine learning to streamline R&D processes [5]. 

Sonita Lontoh , public company board director and former HP's CMO, underscores this potential: 

“AI holds immense promise for advancing sustainability goals, offering solutions for optimizing resource use, enhancing energy efficiency, improving environmental monitoring, and driving innovation across various sectors. From renewable energy integration to supply chain optimization, AI can be a game-changer.”

Yet, for all its promise, AI’s environmental costs cannot be ignored. Training advanced AI models, particularly generative AI (GenAI), requires immense computational power. For example, training a single large language model (LLM) can emit as much carbon as five cars over their lifetimes [6].

Did you know that AI can be surprisingly water-intensive? Or that it contributes to mounting electronic waste? Well, the data centers driving AI operations consume vast amounts of water for cooling, worsening shortages in already water-stressed regions [7, 12]. The hardware underpinning AI, such as GPUs and servers, contributes to the growing wave of electronic waste, with significant environmental consequences from production to disposal [8]. 

Naomi Stone, CESM, ENV SP, CPP , CEO of MugenKioku Corporation and Enso Circular Markets , highlights another critical risk: 

“AI-driven sustainability solutions risk reinforcing existing inequalities if trained on biased data, leading to ineffective or even harmful interventions. Data reliability is also a major concern — many AI models depend on open-source data, which can be outdated, incomplete, or incorrect.”

Robust frameworks and legislation are essential to harness AI’s potential while mitigating its risks. Europe’s landmark EU AI Act includes sustainability considerations, emphasizing the need to minimize environmental impacts during AI development and deployment [9]. The Corporate Sustainability Reporting Directive (CSRD) expanded ESG reporting requirements in 2024. While it focuses on broader organizational impacts, including AI-specific effects is vital. Clear disclosure of AI’s impact strengthens transparency and accountability as organizations increasingly rely on these systems [10]. Meanwhile, in the US, the National Institute of Standards and Technology (NIST) is developing guidelines to assess AI’s sustainability, focusing on energy efficiency and carbon footprint measurement [11]. 

However, gaps remain. Current frameworks lack clarity on measuring and allocating energy usage across the AI supply chain. This underscores the need for standardized metrics and operational methodologies to accurately assess AI’s environmental impact. 

T. Alexander Puutio , author of AI for MBAs and Adjunct Professor at Columbia University , emphasizes the governance challenge: 

“Many sustainability initiatives using AI suffer from either a lack of clear, measurable outcomes or from poor cross-disciplinary collaboration — especially between technologists, sustainability leaders, and frontline implementers. Without mutual understanding, even the best tools fail to drive real impact.”

We’ve already highlighted a few compelling examples of AI driving positive change in the consumer goods sector, but let’s not stop there. There are several other examples of how the sector is quickly changing. In cosmetics, AI empowers companies like Unilever to analyze consumer data and create personalized, sustainable products with minimal environmental impact [13]. In food and beverage, platforms like Too Good To Go use AI to reduce waste by connecting consumers with surplus food from restaurants and grocery stores, which is also a business model innovation [14]. Procter & Gamble , in home care, applies AI to optimize packaging design, cutting material use and waste [15].

All impressive, right? But let’s step back for a moment. While we celebrate these benefits, how often do we pause to consider the other side of AI - the one nobody seems eager to face? The vast computational power driving these innovations comes at a hidden and less well-understood environmental cost. The exponential growth of AI models demands massive data centers, consuming staggering amounts of electricity and water and generating significant carbon emissions. Yet, most AI initiatives in sustainability operate in a vacuum, measuring only their direct gains (...less waste here, greater efficiency there) without accounting for the broader ecological footprint of the technology itself.

If we’re serious about harnessing AI for good, we can’t afford to keep ignoring this blind spot. Innovation without introspection is reckless. For this to truly work, we need to embrace creativity with precision, giving space for innovative solutions while maintaining a strong focus on accurate and rigorous measurement. That means being transparent about AI’s full lifecycle impact and taking a more balanced approach, one that not only celebrates progress but also acknowledges and takes responsibility for any unintended consequences.

Fei Yu 余笑飞 , Managing Director at Alcott Global , offers practical advice for professionals navigating this evolving landscape: 

“Candidates should move beyond AI buzzwords and translate their expertise into practical, industry-relevant applications. Employers want to see how AI can drive real outcomes, whether in optimizing resources, reducing costs, or making informed investment decisions that enhance sustainability.”

Navigating the AI-sustainability paradox requires intentional leadership. Collaboration across industries is essential, as sustainability challenges are complex and interconnected. Equipping teams with the knowledge to implement AI responsibly and sustainably is equally critical. Above all, we must prioritize ethical AI, ensuring transparency, fairness, and accountability in AI systems to build trust and drive long-term sustainability. 

AI holds immense promise for advancing sustainability, but its environmental costs cannot be ignored. By adopting a holistic approach, optimizing energy use, leveraging renewable resources, and embedding sustainability into AI development, we can harness AI’s potential to drive innovation while safeguarding our planet. 

As leaders, we are responsible for ensuring that AI becomes a force for balancing its transformative potential with the urgent need for environmental stewardship.

The conversation doesn’t stop here. Which organizations are truly walking the talk when it comes to AI-powered sustainability? What hurdles are they hitting, and who’s actually nailing the measurement game? Let’s keep the dialogue going!

Together, we can shape a future where AI and sustainability are not competing priorities but interconnected drivers of global progress. 

References: 

1. European Parliament Think Tank, “AI and Climate Change, Artificial intelligence: threats and opportunities” (2023). 

2. Google DeepMind, “AI for Data Center Efficiency” (2023). 

3. L’Oréal, “Sustainable Beauty with AI” Beauty Tech Report (2023). 

4. Nestlé, “AI in Food and Beverage” Annual Review (2023). 

5. Pfizer, “Artificial Intelligence: On a mission to Make Clinical Drug Development Faster and Smarter” AI in Pharmaceuticals (2023). 

6. Heikkilä, M., “AI’s carbon footprint is bigger than you think” MIT Technology Review (2023). 

7. Mytton, D., "Data centre water consumption" Nature (2021).

8. World Economic Forum, "Artificial Intelligence for Efficiency, Sustainability and Inclusivity in TradeTech" (2025).

9. European Commission, “EU Artificial Intelligence Act” (2024). 

10. European Commission, Corporate Sustainability Reporting Directive (CSRD) (2023). 

11. National Institute of Standards and Technology (NIST), "AI Sustainability Framework" Trustworthy & Responsible AI Resource Center (2023). 

12. Azarifar, M. et al.,"Liquid cooling of data centers: A necessity facing challenges" Applied Thermal Engineering, Science Direct (2024).

13. Unilever, "How AI and digital help us innovate faster and smarter" (2023). 

14. Too Good To Go, “AI-Powered End-to-End Solution to help Grocery Retailers Manage their Surplus Food” (2024). 

15. Procter & Gamble, “AI in Home Care - Tackling Plastic Packaging: Sustainable Solutions" (2023).