AI and the Environment
Last week it was revealed that Google’s CO2 emissions have risen by 48% since 2019. The majority of this is due to the increasing energy demand of its data centres, which has been greatly exacerbated by its work on AI. Google are now saying that their goal to reach net zero by 2030 will be “challenging”…
Stories highlighting the environmental impact of these AI tools have been popping up for a while now. Microsoft emissions have seen a similar rise over the last few years, again attributed to AI demand. AI Image generators have been shown to use about as much energy as fully charging a smartphone to create a single image. By 2027, the energy consumed to run AI models is predicted to reach levels comparable to those of countries like Denmark. All of this is without even mentioning the significant increases of water consumption by these data centres.
Many times, it has been touted that we can use AI to save us from climate change, but currently it seems to be digging us into a deeper hole. Experts in the area, including Bill Gates, have told us not to worry; that these energy increases are a small problem, but it will pay off in the long run thanks to productivity and technological improvements.
All of this makes me wonder, is it worth it?
Really, I’d say it depends on your use case. There are some amazing things that AI can be used for, that no previous technology could achieve. But not everything needs AI, there are plenty of tasks and problems that can be solved with much simpler methods, without the massive environmental baggage.
Ultimately, we need to critically think about whether the tools we are using are appropriate for the task at hand. Now that we know the environmental costs, we have to factor that in when deciding if AI tools are the right choice.
