Beyond the screen, artificial intelligence leaves a heavy, hidden footprint on our physical world.

Beyond the screen, artificial intelligence leaves a heavy, hidden footprint on our physical world.
Image: Intellect (generated)

In short: While generative AI promises productivity, it operates as a massive material system. Beyond the electricity required to train models, the technology demands vast amounts of water for cooling and physical infrastructure that strains local ecosystems. We are only beginning to quantify the true environmental cost of our digital queries.

Artificial intelligence is often perceived as an ethereal digital phenomenon, yet it is fundamentally a material system. Its infrastructure—data centers, advanced chips, and cooling systems—requires significant resources, including electricity, water, and land. The rapid expansion of generative AI has accelerated the construction of data centers, which house thousands of servers. These facilities are energy-intensive, with training clusters potentially consuming seven to eight times more power than standard computing workloads. As of 2023, North American data center power requirements surged to 5,341 megawatts, nearly doubling from the previous year.

The environmental impact extends far beyond the initial training phase. Every interaction, such as summarizing an email or generating an image, consumes energy during the inference process. Research indicates that a single ChatGPT query can use five times more electricity than a standard web search. Furthermore, the water required to cool hardware can disrupt local ecosystems and strain municipal supplies. Because the demand for new data centers is growing so quickly, much of the electricity powering them currently originates from fossil fuel-based plants, creating a significant carbon footprint that persists long after a model is deployed.

Measuring these impacts is complex, as results vary based on whether researchers account for hardware manufacturing, idle capacity, or local electricity sources. A 2024 benchmark study highlighted the disparity between tasks, noting that image generation can be significantly more energy-intensive than simple text classification. Ultimately, the environmental burden of AI is a governance and justice challenge. While the benefits of these tools are often global, the costs—such as water withdrawal, land use, and electronic waste—are frequently concentrated in specific communities. Addressing these risks requires a shift toward transparency, efficiency by design, and lifecycle responsibility to ensure innovation does not shift environmental burdens onto vulnerable populations.

Opens the Intellect+ feed with this card first. Swipe up for the next one.

Sources

To cite this card, link to https://intellectplus.co.uk/app/society-environmental-impact-of-ai.

How we source, check and correct cards: Editorial standards.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

Finding something worth knowing…