Introduction
The AI boom has been a dominant topic in mainstream conversations, with a primary focus on its potential impact on our future, including superintelligent AI and job displacement. However, amidst these discussions, we often neglect to address the wider ramifications of our increasing reliance on AI technologies. Just like any other form of consumption, the usage of AI brings forth “externalities” – indirect effects that ripple beyond the immediate realm. In this blog, we shed light on the lesser-explored territory of the environmental footprint and the social consequences of our voracious appetite for AI.
Environmental Impacts of AI
As AI adoption skyrockets, so do its environmental costs. In 2019, the digital technology sector was estimated to produce more carbon emissions than the aviation industry. Although AI’s current carbon contribution is less than 1% of the total, its projected growth suggests a ninefold increase by 2030.
AI tools, such as ChatGPT, are built on complex computational systems called large language models (LLMs), hosted in data centers worldwide. The development, training, and usage of these models costs substantial resources. For example, the carbon footprint of a model similar in complexity to OpenAI’s GPT-3 was equivalent to 60 flights from New York to London. More carbon-intensive grids powering larger models, like GPT-3, exacerbate emissions further.
Moreover, the water demands of data centers, used for cooling servers, raise additional environmental concerns. Data centers, which play a pivotal role in AI operations, require vast amounts of water for efficient cooling, resulting in significant water consumption.
In response to the lack of transparency regarding water footprints, recent studies estimated that training GPT-3 consumed between 210,000 and 700,000 liters of water, akin to producing hundreds of cars. This staggering water usage highlights the substantial environmental impact AI model training can have.
Even a single conversation with ChatGPT was found to “consume” the equivalent of a 500 milliliter bottle of water. Such seemingly insignificant interactions accumulate on a global scale, significantly contributing to AI’s overall water footprint.
The environmental consequences of AI are far-reaching and demand immediate attention and sustainable solutions to mitigate its ecological footprint. As AI technology continues to advance, understanding and addressing its environmental impacts must remain at the forefront of our discussions. Only through proactive measures can we strive for a future where AI innovation coexists harmoniously with environmental preservation.