Continue with Google
Nishtha Baral
In our haste to embrace the technological advancements of artificial intelligence, we risk falling into a familiar and dangerous cycle—one where health and environmental consequences are considered only after significant damage is done. What begins as progress often leaves behind a trail of consequences, recognized only when it's too late. From mechanical simplicity to e-waste complexity, from the rise of cars to the eventual ban on leaded gasoline, history reminds us time and again: innovation outpaces our understanding of long-term impacts.
AI systems are leaving more than just a mark of progress—they're creating a growing mountain of electronic waste and rapidly increasing energy and water consumption. We now stand at a critical crossroads. The environmental footprint of the digital revolution is expanding at an alarming rate. By 2030, AI data centers are projected to consume 4.5% of global energy production, use water equivalent to that consumed by entire nations, and contribute to hardware obsolescence that could result in an estimated 5 million metric tons of AI-related e-waste annually—a potential 1,000-fold increase from 2023.
Unlike past transitions, the pace of AI advancement offers little time for analysis or mitigation. Drawing on lessons from history—such as the decades-long health crisis caused by leaded gasoline—this article explores the environmental costs of an AI-driven future. Understanding how past technologies led to environmental and health disasters helps us recognize early warning signs and act before crises become irreversible.
Consider the evolution of automobiles. First introduced in 1886 by Karl Benz, cars initially had no electronic components. Over the next century, electronic systems were gradually introduced, culminating in the rise of electric vehicles (EVs). From virtually none in 2010 to 26 million EVs by 2022, the growth has been exponential. In 2022 alone, EV sales surged by 60%, making up over 14% of global car sales. However, roughly 70% of EV batteries and components are made in China, a country heavily reliant on coal. Battery manufacturing for EVs is significantly less efficient than for conventional vehicles. Improper disposal of these batteries can release toxic materials, exacerbating environmental costs. It took over a century for the automobile industry’s e-waste burden to become evident. In contrast, AI and robotics are accelerating toward that same crisis within a few short years.
“The Secret” author Rhonda Byrne has new release in November
In the 1920s, General Motors engineer Thomas Midgley Jr. introduced tetraethyl lead (TEL) to reduce engine knock, unaware of its toxic legacy. By the 1960s and 1970s, the health impacts were undeniable—lead poisoning linked to premature deaths, cognitive deficits, and increased crime. Despite this, leaded gasoline was only globally phased out in 2021, decades after the first bans. A similar story may unfold with AI if action is not taken promptly.
Modern AI systems—particularly large-scale models—require power-intensive hardware such as GPUs, CPUs, and memory modules. These become e-waste filled with valuable metals like gold, silver, and rare earth elements, alongside hazardous materials like lead and mercury. Unsafe disposal practices in developing countries often involve open-air burning and acid baths, releasing toxins that can cause cancers, miscarriages, neurological damage, and more.
“It’s far easier and more cost-effective to address the e-waste challenges posed by AI now, before they escalate beyond control,” says Asaf Tzachor, a researcher at Reichman University in Israel. His advice highlights the urgency of acting now, before AI causes irreversible harm.
AI’s environmental impact doesn’t stop with hardware. Training AI models consumes vast resources. For example, one generative AI search uses four to five times more energy than a standard web search. OpenAI CEO Sam Altman warned at Davos that AI is heading toward an energy crisis. Already, systems like ChatGPT are estimated to use energy equivalent to 33,000 homes.
Cooling systems for AI servers require enormous amounts of water. In July 2022, a cluster serving GPT-4 consumed about 6% of its district’s water. Water use by Google’s Bard and Microsoft’s Bing spiked by 20% and 34%, respectively, in one year. By 2027, global water demand for AI may rival that of the entire UK. These are not just numbers—they’re wake-up calls. The tools we praise for their intelligence are quietly consuming resources at a scale comparable to entire nations. It’s unsettling to think that the same technology we rely on so heavily may be undermining our environmental sustainability. Meanwhile, regulatory responses remain weak. Governments must implement stronger frameworks—mandating responsible hardware disposal, capping energy use in data centers, and enforcing accountability for e-waste.
The UN Sustainable Development Goal 12 calls for sustainable consumption and production, but e-waste increased by 38% from 2010 to 2019. Governments, manufacturers, and consumers must heed this warning. The UN Environment Programme (UNEP) recommends five key actions to address the environmental impact of artificial intelligence. These include establishing standardized procedures to measure AI’s environmental footprint and requiring companies to disclose the environmental costs of their AI-based products and services. UNEP also urges improving the efficiency of AI algorithms, promoting the recycling and reuse of components, and encouraging the greening of data centers through the use of renewable energy and carbon offsetting. Finally, it emphasizes the importance of integrating AI-specific regulations into broader national environmental policies to ensure a cohesive and sustainable approach.
But where does responsibility truly lie? Are companies solely to blame—or are we, as eager consumers, also accountable? Our desire to be early adopters feeds the demand for constant upgrades. Especially among youth, owning outdated gadgets is seen as unfashionable. This mindset drives unsustainable consumption. Sustainable innovation demands more than green manufacturing—it requires a cultural shift. Conscious consumerism must become the norm, not the exception. Too often, sustainability enters the conversation only after damage is done. We need to change that.
In Nepal, AI and robotics are beginning to play roles in disaster management and agriculture. As the nation takes its first steps into this field, it must learn from the missteps of more advanced nations. With the right policies and foresight, Nepal can avoid repeating the same mistakes. Failing to act now could result in serious consequences in the near future.
The author, a core member of RedPaper, crafted this article during her time at RedPaper, with guidance on exploring a topic that intrigued her.
