AI and Climate Change: How Digital Trends Are Harming Our Planet

Poornima

The Industrial Revolution of the 16th century was a complete game changer. The result was a tremendous advancement in technology, which led to the transformation of economies and industries and ultimately made life easier in many respects. But it also came with a cost; however, it was the rise of pollution, deforestation, and climate change—issues we still grapple with today.

In the 21st century, we are again going through a massive transformation—the rise of Artificial Intelligence (AI). AI has the potential to revolutionise industries, but as with previous transformations- AI has some unintended side effects. One such area, already seeing the impact of AI, is in AI-generated art (notably in the recent wave of digital recreations of the iconic Studio Ghibli art style).

The Rise of AI-Generated Ghibli Art

AI tools such as DALL·E, Midjourney and Stable Diffusion make it possible for anyone to create beautiful Ghibli-style images in seconds. It’s an exciting and creative development that has rapidly evolved online. But behind every AI-image is a considerable amount of processing power, which translates to significant energy use, and energy use isn’t free; energy use produces carbon emissions, which feeds a burdening climate event.

While AI-generated art seems fun and non-threatening, we should recognise the reality: even the art we generate digitally has an environmental impact.

The Environmental Cost of AI-Generated Art

AI-generated pictures can appear to be harmless digital entertainment, but the situation is more complicated than that. The computation behind generating these images is heavy, depending on large data centers that consume vast quantities of electricity and water to operate. These centers produce large carbon footprints, which lead to global warming.

Carbon Footprint and Energy Consumption

AI art creation is computationally intensive and hence energy-intensive. As per studies:

It takes about 11.49 kilowatt-hours (kWh) of energy to create 1,000 images with Stable Diffusion's XL 1.0 base model, releasing about 1,594 grams of CO₂, equivalent to driving a gasoline car for approximately 4.1 miles.

● To train a single large AI model like GPT-3, 1,287 megawatt-hours (MWh) of electricity was used, emitting the same as 502 return trips by air from New York to San Francisco.

● Google estimates that inference (the activity of applying an existing trained model to produce new output) contributes 60% of AI energy consumption, which is more than electricity consumed in training.

Water Consumption

Another underappreciated environmental effect of AI is its wasteful consumption of water. Data centers utilise huge quantities of water for cooling purposes, resulting in a resource shortage in places already lacking water. For example:

● Google's data centers consumed 4.3 billion gallons of water in 2021 alone to keep from overheating.

● AI tools such as ChatGPT may consume up to 500 milliliters of water per basic AI engagement.

E-Waste and Hardware Disposal

AI tech relies on powerful GPUs and processors, which render them obsolete almost as soon as they are manufactured due to swift advances. This means more e-waste, with old hardware ending up in the trash. Many devices use dangerous materials that, when dumped carelessly, can pollute the environment and create health issues.

The Climate Change Connection

Artificial intelligence art production drives environmental expenses that connect directly to climate change phenomena. The ongoing need for AI-produced material results in:

● The release of greenhouse gases intensifies global warming at an accelerating rate.

● Power grids dependent on fossil fuels face increased stress due to rising energy consumption.

● The exhaustion of essential resources such as water and rare earth minerals critical for computing hardware production.

The United Nations Environment Programme (UNEP) states that to keep the global temperature increase within 1 degree, carbon emissions need a 45% reduction by 2030. 5°C above pre-industrial levels. The exponential increase in AI applications may cause data center emissions to reach three times their current levels before the decade concludes.

Expert Warnings

Environmental experts are sounding the alarm about AI's increasing energy needs:

“AI’s environmental footprint is often overlooked, yet it plays a crucial role in accelerating carbon emissions. We need better transparency and regulation in AI energy use.”— Sasha Luccioni, Climate and AI Lead at Hugging Face (Time Magazine)

AI and Creativity: A Threat to Human Artists?

Aside from its environmental concerns, AI-generated art threatens human creativity. Studio Ghibli's hand-drawn animation is valued for its emotional depth and craftsmanship. AI-generated versions can replicate the appearance but lack the soul, uniqueness, and cultural significance of genuine human-made art.

Moreover, AI-generated content:

● Limits career opportunities for illustrators and animators.

● Undermines creative originality, as AI models replicate and adapt existing art.

● Raises ethical questions about copyright and intellectual property.

As per a World Economic Forum report, AI may displace 85 million jobs by 2025, most of them in the creative industry.

Global Efforts for Sustainable AI

The global community is aware of the environmental consequences of AI. International conferences such as IAIA25, ICAIS, the Global Conference on Energy and AI, and Sustainable AI 2025 all desire to address AI's responsible development, reduce AI's carbon impact, and promote sustainability in technology as a global initiative that engages all stakeholders.

Balancing AI Innovation with Environmental Responsibility

The Industrial Revolution created amazing advances, but with unwanted environmental degradation. Now, with AI, we have the same challenge. Whereas AI-generated art in the style of Ghibli and other digital phenomena brings convenience and ingenuity, they have an environmental price that can't be ignored.

As we go ahead, we need to ensure that we prioritise sustainability, keep AI in line with climate objectives, and also promote human creativity whose work cannot be done by machines. By choosing wisely, we can reap the dividends of AI while safeguarding our earth for generations to come.

"The greatest threat to our planet is the belief that someone else will save it."

— Robert Swan, Environmentalist