Artificial Intelligence (AI) presents both environmental challenges and opportunities. How we use AI will determine whether the tool will help or harm humanity.
The Negatives
AI has a large energy footprint. The training process for a single AI model, such as a large language model (LLM), can consume thousands of megawatt hours of electricity and emit hundreds of tons of carbon.
An AI training model can lead to the evaporation of a significant amount of freshwater into the atmosphere for data center heat rejection, potentially exacerbating stress on limited freshwater resources.
There is also an uneven distribution of how different regions and communities are impacted by AI, with disadvantaged regions often bearing the brunt of negative consequences.
The data centers that we use for all our computing needs produce from 2.5 to 3.7 percent of global greenhouse gas emissions. That will increase with more AI if the energy that powers AI is not green.
More AI will mean more energy consumption. Estimates show that AI use in Europe is expected to grow 28 percent by 2030. You see similar numbers for the rest of the developed world. In the US, according to a new report released by the Electric Power Research Institute (EPRI) data centers that power AI models could account for up to 9.1% of the country’s overall energy demand by the end of the decade.
Companies and policymakers need to consider the environmental impacts AI may have. Does AI lead to an increase in GHG emissions, or can AI be powered by renewable energy sources? Is the water used for cooling AI training systems readily available, or does it unnecessarily strain local water resources to the detriment of other stakeholders?
The Positives
The promise of AI is that it can aid us do jobs better and more efficiently. This includes things like climate modeling, where AI can help scientists better understand climate patterns, identify trends, and make predictions related to climate change. Scientists hope that AI can also help develop climate mitigation strategies.
AI can also help with things like water conservation by better managing water resources, and wildfire prevention and detection, by better predicting literal hotspots where fires may start.
AI can be used to address environmental inequality by helping scientists, businesses and policymakers prioritize solutions for disadvantaged regions.
AI may also lead to more leisure. In 1930, economist John Maynord Keynes famously theorized that we would be only working 15 hours a week by the turn of the century due to increases in efficiency. If AI can help us do our jobs more efficiently we could work fewer hours.
In conclusion, AI is just a tool, like any other tool. It can be used for good or for ill. It can be policed and regulated so that energy use and outcomes can be monitored, or it can be lightly policed and decisions left up to those running AI systems and hoping for the best.
Past experience argues for a balance that can mitigate harm and accentuate positive outcomes, but that is always a difficult balance to strike.
To whom is this relevant?
Companies looking to use AI as a tool to improve their operations.
Investors looking for best practices in the use of AI.
Policymakers charged with setting regulations for AI.