Environmental concerns

GAI tools (as for all AI tools) need a lot of computing power, which in turn means a lot of cooling of computer equipment, which uses a lot of water (Ekin, 2019), as well as increased environmental impact from the manufacture and disposal of the computers themselves. It is difficult to calculate the environmental impact of training the models and making individual requests, because, you’ve guessed it, of the lack of transparency in the processes. The Organisation for Economic Cooperation and Development (OECD) has recommended a set of approaches to measure and monitor the costs of AI (OECD, 2022), to help governments with policy-making. I'm not aware of any current work happening in this area, and it may be that we need to try doing this work at a local level. Bashir et al (2024) make a really good case for systematic and targeted development of GenAI so that development is worthwhile, and not just novel. 

Whilst this section is about concerns with AI, it is worth noting that many environmental researchers are enthusiastic about the use of AI tools in general (not just generative ones) to help track changes and predict the effect of interventions, so we have to find a balance here. The UK newspaper The Guardian recently quoted Sasha Luccioni, a researcher in ethical and sustainable AI at a US tech company, Hugging Face, saying: “Fundamentally speaking, if you do want to save the planet with AI, you have to consider also the environmental footprint [of AI first…It doesn’t make sense to burn a forest and then use AI to track deforestation.”  (Luccioni, quoted in Stokel-Walker, 2023)

If you would like to know more, why not listen to this podcast episode from the IIEEE at Lund University, on Exploring the role of AI in sustainability ? Links to an external site.

In September 2023, a preprint suggested that "The Carbon Emissions of Writing and Illustrating Are Lower for AI than for Humans Links to an external site." (Tomlinson et al, 2023), with a  comparison of the energy needs of humans writing, using computers, with AI generating text. There are some major flaws with this, by the way, since we breathe and consume food and water when we aren't writing, but as a general point, more research is definitely needed! 

Summary: we should consider the benefits and challenges of AI from an environmental as well as a societal perspective.