[Changemaker Series] AI: A Tool for Change or a Power Drain? - Unpacking Our Conversation With Geraldine de Bastion

This is the next part of our Changemaker Blog series.

Here, we share insights from our conversation with Geraldine de Bastion.

Geraldine is a digital rights advocate, researcher, and policy expert working at the intersection of sustainability, digital infrastructure, and human rights. Geraldine is a member of the executive board at Global Innovation Gathering and has been deeply involved in work that seeks to reimagine the internet as a more equitable and sustainable space. 

This conversation provided interesting insights into sustainable technology and the energy consumption of AI. Geraldine offers candid insights and examples that challenge the dominant narratives around innovation. Our conversation involved reflections and hopeful alternatives for anyone looking to understand the ecological cost of the internet and what we can do about it. 

We are stoked to share these insights through this blog to inform the larger dialogue around sustainable tech futures. 

The Complexity of Sustainable Internet Infrastructures

What are some of the system-level challenges in making the internet and emerging technologies like AI more sustainable? Where are the biggest gaps - in awareness, policy, research, or industry?

Geraldine:

This is more a collection of reflections than a structured assessment, but one of the biggest issues is the lack of informed discourse about what sustainability actually means in the digital space. For example, even basic data, like temperature impacts of server operations or hardware longevity, is fragmented. Some approaches focus on short term electricity consumption, while others value long term reuse of hardware. Without systematic data and shared definitions, policymakers and technologists are navigating in the dark.

We also see infrastructure inefficiencies due to poor regulation. Take Europe: some countries like those in Scandinavia mandate telco tower sharing, while others allow every company to build their own infrastructure - that is duplication and waste. On the software and AI front, we are in a race to build the biggest models instead of collaborating through open source projects. Why are we all feeding massive datasets into LLMs repeatedly when initiatives like Hugging Face show we can do it differently?

For the understanding of our readers, Hugging Face functions as a collaborative environment where developers and researchers can share, discover, and utilize a vast collection of AI resources, making advanced AI more accessible and open. It provides users with access to pre-trained models, datasets, and tools for building, training, and deploying AI applications, particularly in Natural Language Processing (NLP). 

As the conversation delved deeper into AI, Geraldine reminded us of the ecological cost hidden behind the digital tools we use every day.

“People were shocked when I told them that generating a single picture with AI might use as much energy as charging your phone,” Geraldine said. “We are encouraged to play with these tools, but rarely do we discuss the energy behind them.”

This conversation also enlightened us with a subtle anecdote about the functioning of different AI chatbots we have gotten so used to talking to. Geraldine chuckled at OpenAI CEO Sam Altman’s suggestion that we stop saying “please” and “thank you” to AI because it uses more computational power, while at the same time, AI tools like ChatGPT are trained to compliment users with emojis and polite phrases. We are not designing these systems for sustainability, yet we place the burden of responsible use on individuals.

Building a Toolkit for Youth

What should be included in a toolkit aimed at helping young people understand the internet’s energy consumption? How can they meaningfully contribute?

Geraldine:

I think awareness is crucial, but we also need to be careful about overburdening individuals, especially young people. Coming from Germany, I am used to navigating complex individual choices like sorting trash that requires citizens to separate trash into color-coded bins for recycling (plastics, paper, organics, residual waste, and glass). We often think about what cars we use, whether we would want to use petrol, diesel, or Compressed Natural Gas in our vehicles. But I also think this hyperfocus on individual action often hides systemic accountability.

Today’s Gen Z and Gen Alpha are inheriting a planet full of climate crises, and we are also telling them they are responsible for fixing it. That can be overwhelming.

A good toolkit would:

• Help youth understand their individual power

• Provide advocacy tools to create systemic change

• Highlight alternatives to Big Tech. Geraldine refers to these alternatives as small tech. 

There are powerful, ethical alternatives out there - they might not be as popular, but they offer pluralistic, sustainable digital futures.

AI: A Tool for Change or a Power Drain?

Do you think AI can help make the internet more sustainable? Or is it contributing to the problem?

Geraldine:

It is hard to stay entirely optimistic about the environmental promise of AI right now. As much as generative AI tools are being celebrated for their potential to accelerate sustainable innovation, their current reality feels far from green. The sheer scale of computing power, the number of data centers being established, and the renewed reliance on atomic power and fossil fuels to keep them running have turned AI into a significant energy sink.

Recent insights from the International Energy Agency also highlight that the electricity consumption tied to AI is projected to rise steeply. While AI might help us map and manage sustainability challenges, the systems enabling it are themselves unsustainable. To consider this: we are trying to use AI to address the very problems it exacerbates.

When we are developing a youth-focused sustainability toolkit, this dilemma is even more evident. We do not want to present a dystopian future or feed into tech pessimism, but rather help young people see how they can make mindful, small choices - even something as simple as preferring a traditional web search over a heavy AI query - while also understanding the larger systemic issues at play. After all, much of the problem lies in profit maximization models that prioritize growth and performance over long-term ecological balance. Can you share some examples of initiatives that you have come across that are working towards making the internet more sustainable?

Geraldine:

In this context, some of the most inspiring examples come from outside the mainstream tech hubs. I recently spoke with a South American activist who works at the intersection of digital rights, climate justice, and community empowerment. Her work reimagines how digital infrastructure development can align with climate protection and local rights, offering a grounded, human-centered alternative to the extractive model that currently dominates.

These kinds of initiatives remind us that sustainability is not just about reducing data center emissions; it involves rethinking who technology serves, how it is built, and what kinds of futures it enables.

Real Life Initiatives and Examples

What are the most important, but often overlooked, aspects of the internet’s ecological cost that young people should know? Can you share examples of promising initiatives making the internet more sustainable? 

Geraldine:

Absolutely. There is brilliant work being done. 

When we think about AI and sustainability, it is easy to get stuck in debates about energy consumption and environmental impact. But there is another side that is imaginative, practical, and inspiring. 

In Germany, the Ministry of Environment is exploring AI-powered platforms that can guide users through everyday decisions: 

Should I fix this gadget? Donate it? Take it apart to reuse parts? 

By analyzing options and connecting people with local repair shops or donation networks, AI can make sustainability actionable at the community level.

Other interesting projects involve great innovation. 

One example is Project Cetacean Translation Initiative (CETI), a research effort using AI to decode whale communication. Imagine having a conversation with another species! While the project raises ethical questions like could we misinterpret or exploit these interactions? - It also opens up a radical idea of AI as a bridge between humans and the natural world, helping us reconnect with the planet in ways previously unimaginable. 

There are also applications like Flora Incognita developed by institutions like the Technical University of Ilmenau and the Max Planck Institute for Biogeochemistry. Through this AI-powered mobile app, one can identify wild plants from photos. The app provides information on the characteristics, distribution, and conservation status of the plant. This application can help in research and track overall conservation patterns. 

Beyond these examples, AI is already transforming fields like medical research and environmental science by handling vast amounts of data and revealing patterns humans alone could never detect. The key, as experts caution, is to balance excitement with reflection. For every incredible application, we need to remain mindful of ethical, social, and environmental considerations.

We are deeply grateful to Geraldine for sharing these insights. These insights provide a clear and thoughtful perspective that balances doubts about today’s technology with hope in collaborative, innovative, and community-driven solutions. It confirms that the ecological cost of the internet is a policy and design challenge, not just a problem for individual users to fix. However, it is only with collaborative effort that we can move towards more sustainable tech solutions.