From Eco-Anxiety to Action: How AI Can Drive a Sustainable Future

As a Product Owner, my days are a whirlwind of innovation. Yet, beneath the surface, a growing concern whispers: eco-anxiety. This isn’t just a personal feeling. It is a global challenge exacerbated by the very technology we use in our daily lives – Artificial Intelligence. The good news? You can use AI for a sustainable future. It has been redefining our carbon neutrality targets, though not always in obvious ways.

Redefining Carbon Neutrality Targets with AI and the Awakening of Consciousness

The journey to carbon neutrality is complex, and with the rise of AI, it has become even more challenging. We are not all yet fully aware of the profound changes and new responsibilities that this emerging technology imposes on us. Luckily, it is just a matter of time.

The Environmental Impact of AI: Shall We Continue to Ignore an Undeniable Reality?

AI’s environmental footprint is significant, extending across the entire lifecycle of its systems, from development to implementation and continuous maintenance. Let’s look at some of the main areas that are impacted:

Energy Consumption

Energy consumption is one of AI’s biggest challenges. Given their complexity, AI models demand substantial computational power. First, the training phase for sophisticated models can be extensive and energy-intensive. The energy used for inference (the ongoing use of an already trained model) is typically far greater than the energy consumed during its initial training. Deep learning, natural language processing (NLP), and Generative AI (GenAI) models, for instance, are more energy-intensive than simpler classification models.

The graph provided by EY illustrates the substantial CO2 equivalent emissions for training various LLMs. This reflects the increasing complexity and computational demands of these models, with parameters growing from 100 million to 500 billion.

Greenhouse Gas Emissions

Greenhouse gas (GHG) emissions are a direct consequence of AI’s high energy consumption. The significant energy required to power AI systems often leads to considerable GHG emissions, especially if the energy is sourced from non-renewable power plants.

Water

Water consumption is a factor often overlooked. Large data centres – essential for training and implementing increasingly advanced AI models – require cooling systems that consume vast amounts of water to prevent overheating and maintain performance. This use can exacerbate water scarcity in regions already vulnerable to such stress.

E-waste

Finally, hardware and electronic waste represent another environmental impact. The hardware supporting AI, including servers, GPUs, and other specialised equipment, has a significant environmental impact. The production and disposal of these devices involve substantial resource extraction, manufacturing emissions, and electronic waste, which can lead to environmental degradation and pollution.

Avoid or Force it to Be Part of the Solution?

As a Product Owner, I try to alleviate my eco-anxiety by balancing the urgency of innovation with the responsibility of developing more sustainable digital solutions. It is a constant struggle to ensure that the products we build do not significantly contribute to the climate crisis, but rather, mitigate it. Nowadays, we can analyse the energy consumption of LLMs through data.

And this knowledge is what drives me to seek solutions and to ask: how can we do better? First and foremost, it is crucial not to succumb to anxiety, but to transform it into action. Awareness of the problem is the first step towards change, and AI, paradoxically, can be part of the solution.

Can AI Be a Catalyst for Sustainability?

Despite the inherent challenges of its environmental footprint, AI has the potential to accelerate sustainability – and at this juncture, with the climate crisis upon us, time is something we do not have in abundance. Studies show that when used seriously and intentionally, AI can significantly improve our carbon neutrality targets, including having direct impacts on the dynamics of IT teams.

A Good Initiative: AI for Good

The “AI for Good” movement, driven by organisations such as the United Nations, demonstrates how AI can be leveraged to achieve the Sustainable Development Goals (SDGs). For example, it is estimated that by 2030, AI could be used to reduce global GHG emissions by 1.5% to 4%.

A recent World Economic Forum report highlights how AI is revolutionising global trade, driving unprecedented levels of efficiency, sustainability, and inclusivity. AI has the ability to “simplify complexity”, such as:

• Improving Supply Chain Visibility: By integrating real-time data, AI allows for rapid anticipation of changes and prediction of disruptions, which is vital for compliance and resilience.
• Promoting the Circular Economy: AI accelerates the transition to circular models, prioritising resource conservation and reuse. AI and blockchain tools track product lifecycles and identify recycling opportunities, as seen with electric vehicle batteries.
• Reducing Scope 3 Emissions: AI offers insights to manage value chain emissions (Scope 3), which can represent up to 95% of an organisation’s environmental footprint.
• Optimising Logistics and Operations: In logistics, AI uses predictive analytics and real-time data simulations to plan routes, forecast demand, and automate tasks, reducing errors and freeing up human resources for strategic activities.
• Facilitating Trade for SMEs: AI can support small and medium-sized enterprises (SMEs) in navigating the complexities of global trade, simplifying agreements and assisting with compliance.

Transforming Anxiety into Fuel for Change

As a Product Owner, being an integral part of this journey is both a privilege and a responsibility. My role can be decisive in guiding teams and projects to a successful outcome, by utilising the right AI tools and advising clients on the best digital solutions, always focusing on their objectives without neglecting sustainability.

Daily Practice Yields Long-Term Results

In my day-to-day work, advocating for sustainability in AI begins with product planning. This means challenging teams to consider not only functionality and performance, but also the environmental impact of the solutions we are developing. It is essential to incorporate sustainability considerations at every stage of the AI systems’ lifecycle, from development to implementation and monitoring.

Collaboration is fundamental. It is crucial to work as a team, with diverse perspectives and backgrounds, including designers, engineers, data scientists, and even sustainability specialists, to ensure environmental concerns are addressed proactively. The goal is for sustainability to become an inherent feature of the product, rather than merely an afterthought.

Above all, we must be critical: Is AI truly necessary for this particular need? Is it worth the effort, or are there less costly alternatives for the environment, the business, and the customer?

It is Undeniable that the Sustainable Future of AI is in Our Hands

AI is a double-edged sword. On one hand, it drives efficiency and innovation, offering solutions to pressing environmental challenges. On the other, its exponential growth and the increasing complexity of its systems exacerbate the environmental challenges inherent in technological adoption.

Although legislative frameworks are beginning to emerge to integrate sustainability considerations into the development, operation, and use of AI, significant gaps persist that hinder organisations’ practical compliance with these standards. The absence of universally accepted measurement techniques makes it difficult for organisations to accurately assess and communicate their sustainability efforts.

To advance the intersection between AI and sustainability, a more holistic approach to measuring AI’s environmental impacts is essential, considering compute operations, electricity usage, carbon footprint, and water usage.

As a Product Owner, and also a mother, my eco-anxiety is a driving force. It is a constant reminder that the future of AI, and the future of our planet, are in our hands. By embracing AI with intentionality, transparency, and an unwavering commitment to sustainability, we can transform this powerful technology into a catalyst for a more equitable, efficient, and, above all, sustainable world.