The following is a transcript of the video above, from our premium webinar “AI for An Equitable World: How Your Organization Can Approach AI Responsibly.” If you’d like to access the full webinar, become a Leading Edge member. As a Leading Edge member, you’ll gain access to 60+ webinars and our full library of digital content with essential insights for purpose-driven leaders.

We are then running that through data centers that in other parts of the world are using over 20 ounces of clean water to cool the system to test if that person can drink water where they are.”

Amy Sample Ward: The environmental impact of AI is really complicated, because for some folks, and very validly, we can say this isn’t new. Technology was already harming communities and environment. It already took a bunch of energy to run the internet….Technology already had harmful environmental impacts, And AI is using water and electricity and associated resources, including just physical buildings, at a rate that is honestly difficult for scientists and researchers to even calculate. So, yes, it is a tool….There is an ironic catch-22 right? There have already been AI-enabled technologies supporting environmental tracking, weather prediction, and in order to run those tools somewhere else in the world is just being destroyed.

So there’s an example of teenager, young tech innovator…and they live in Michigan, where, as we know we’re going on however many years now, without clean drinking water. And so they built an AI tool off of Gemini, I think, that community members could use to test if their water was drinkable. Okay, so we are then running that through data centers that in other parts of the world are using over 20 ounces of clean water to cool the system to test if that person can drink water where they are. And I’m not saying any of it is right or wrong, because we are not living in a world where we have the luxury of binaries like that. Okay, this is a really complicated issue, and so it’s not okay [to say] don’t use any AI tools because they are so environmentally resource strained. It is to know that there is that strain, to demand that if we’re going to use a product, that that tech company or provider is aware of their own environmental impact, is willing to talk about it transparently and what they’re doing about it. This is not a buy-a-tree offset situation.

And if our demands, if our requests for having that information from tech companies are or are not coming, that’s one layer. But then it’s also for us to say, hey, we know that using something like ChatGPT is really resource heavy. We are not going to use it every day, 100 times. We’re going to use it when we want to use it and really need to use it, versus, “Oh, let me just do this 18 times.” So knowing, again, what’s right for you. That answer can be different and will be different for every single person on the call. But how do we say, we have our eyes open to these environmental impacts, and this is how we want to move through this, knowing what those impacts are.