Artificial intelligence (AI) is becoming more integrated into our everyday lives. Unfortunately, as it does, the environmental impact of AI is becoming a huge red flag. Experts are increasingly calling attention to the fact that although AI vastly increases efficiency and convenience, it’s a trade-off with massive energy consumption and resource demands. This ambivalence has led people like artist Jon Ippolito to avoid using AI whenever possible. He drives home an important lesson: we all need to be aware of its hidden, pernicious costs.
Ippolito, also a proponent of more mindful, intentional use of technology, pointed to some shocking statistics on technology’s energy use. According to his app, a simple AI prompt such as “Tell me the capital of France” consumes 23 times more energy than a conventional Google search devoid of AI features. Moreover, a complex request like “Tell me the number of gummy bears that could fit in the Pacific Ocean” amplifies that consumption to 210 times more energy.
Since then, innovation manager Marissa Loewen has primarily deployed AI as a project management tool within the 2014 Band TMA. She is aware of the ecological footprint of her tech decisions. “We try to use it conscientiously though because we do realize that there is an impact on the environment,” she stated. Loewen’s presentation emphasized the advantages of local, noncloud computing alternatives. These digital solutions give transportation agencies deeper control over how and when they are using energy and managing data.
The environmental cost doesn’t stop with easy writing prompts. Sasha Luccioni, AI & Climate Lead at Hugging Face noted it well. Creating a single, high-definition image consumes roughly the same amount of energy as charging a smartphone for 47 percent of one full charge. Her work has already demonstrated that streaming Netflix for an hour uses more energy than generating a complicated AI text prompt. In reality, a Zoom call with only 10 people consumes 10 times more energy than asking that same prompt!
Noman Bashir, computing and climate impact fellow at the Massachusetts Institute of Technology’s Climate and Sustainability Consortium. He deeply amplified these concerns. He noted that the buildout of data centers is rapidly accelerating. They typically rely on fossil fuels since we currently are not able to put enough renewable energy onto the grid. “Since we are trying to build data centers at a pace where we cannot integrate more renewable energy resources into the grid, most of the new data centers are being powered by fossil fuels,” Bashir explained.
The amount of energy these bigger data centers consume is just mind-blowing. Some facilities even consume a jaw-dropping 5 million gallons (18.9 million liters) of water daily. That’s more than enough to power the average daily needs of a town of 50,000 people! These numbers are like leaving an incandescent lightbulb on for more than a year. While dystopian in nature, they nonetheless pose thought-provoking queries about the tech industry’s sustainability.
Ippolito advocates for minimizing reliance on data centers, stating, “If you can do anything that cuts a data center out of the equation, I think that’s a win.” His viewpoint highlights the need to seek out different solutions that achieve the same outcome, an increase in technology and innovation, but with a decreased consumption of energy.
Luccioni raised a key point that further complicates the generative AI debate: The generative AI process is fundamentally different than traditional processes. Instead of building on the often excellent work done across the country, it’s completely re-doing the wheel. And that requires a lot more compute,” she added. This serves to confuse the narrative around the relative efficiency of AI and the environmental costs associated with the technology.
Even with these challenges, pioneers are making big moves in the direction of greener AI practices. For instance, Luccioni utilizes Ecosia, a search engine that employs efficient algorithms and invests profits in tree-planting initiatives to mitigate environmental impacts. These decisions may stand as an initial indication that users are becoming more conscious of the environmental impact of their online actions.
Ippolito cautioned against complacency regarding technology’s environmental impact: “It’s not just about making people conscious of AI’s impact, but all of these digital activities we take for granted.” He draws parallels between past technological shifts and current trends, stating, “When cars replaced horses, commute times didn’t shrink. We just traveled farther.” This amusing comparison is a cautionary fable for our age, that progress is always accompanied by a host of unintended side-effects.
