Generative AI: Environmental Considerations

This week in class, we continued discussing generative AI and its implications in our world today. Specifically, we discussed the negative environmental impacts of AI and what options we have to mitigate these effects. Do we have a choice in using GenAI in today’s society? Are there ways to avoid using AI? Does our AI usage even make a difference in the grand scheme of things? These are all questions we discussed this week and had the opportunity to learn more about.

When it comes to AI, do we have a choice in using it?

AI is ingrained in society, and the rapid increase of online tools (websites, search engines, software, and generative AI sites) that utilize AI has led to widespread usage of these systems. Even if we are not using it directly, it is challenging to avoid it altogether, and we often still end up interacting with systems that use AI in some way. Avoiding the use of AI would be a significant challenge, and I would argue that the more important challenge is learning how to critically analyze (and be able to identify) AI outputs, and be intentional about our AI usage.

Let’s talk about the environment…

The field of research regarding AI and its environmental impact is steadily growing. Until recently, the literature on this topic was relatively limited. However, emerging research has indicated that AI poses environmental concerns related to water and electricity usage. According to one website, the annual global impact of AI is anticipated to account for more water withdrawal than half of the United Kingdom. Additionally, two-thirds of the time, AI data centers are built in locations already suffering from water scarcity. Concerning electricity, research has indicated that the rise in AI will cause energy systems to struggle to keep up with demand, and energy consumption will only continue to grow. Both energy and water consumption for AI systems could lead to greater impacts on the climate, as well as the misallocation of resources if energy is provided for AI systems instead of essential social services (such as healthcare).

There are various ideas about how to curb AI’s impact on the environment, one of which is Elon Musk’s orbital data centre solution. While Musk (and other tech giants) have proposed potential solutions, it remains to be seen if these are concrete possibilities or simply talk. As discussed in class, I would agree that the primary goal of these tech giants is not to save our environment, but rather to maximize their own profit and influence. Other potential solutions include instituting a carbon tax for companies using AI software or systems: AI models would be taxed based on their usage, and the money would be directed towards sustainable energy solutions. Lastly, another solution is to solve the problem of AI using AI. Many people believe that AI can be used to find a solution to the climate crisis by enhancing energy efficiency and fostering innovation in this field. This remains to be seen.

What’s next?

This week pushed me to critically examine my own relationship with AI usage. I think that AI is a valuable tool, but it is also important to consider the environmental impact you have when you use AI systems in your daily life. While it feels impossible to completely avoid using AI, I think this week’s discussions pushed me to be more intentional in my use of AI. Learning about the real-world consequences of my AI use has made it feel less abstract, and I recognize that my actions have an impact. However, in the grand scheme of global AI usage, it is also important to recognize that my decision to avoid AI use is a drop in the bucket, and the increase in AI integration is a global movement. Moving forward, I think it is important to be informed about AI and use it critically, not habitually. Although my own awareness is just one piece of a very large puzzle, it is a step in the right direction.