What does AI mean for Sustainability?
Any discussion of Artificial Intelligence is incomplete without due consideration of its impact on sustainability
Given the significant implications of AI for sustainability and environment, in our July AI Ethics Twitter Chat, we invited Dr. Dorothea Baur Ethics Officer, Lecturer, Author (@DorotheaBaur) to discuss “What does AI mean for Sustainability?”
Mia Dand: Welcome Dr. Dorothea Baur! Thank you for joining us. Let’s start with the basics. What does sustainability mean?
Dr.Dorothea Baur: Thanks for having me! The standard definition of sustainability stems from the Brundtland Report which was published in 1987 and which defined sustainable development as one that satisfies the needs of the present without adversely affecting the conditions for future generations. In general, sustainability is thought to encompass well-being along three dimensions — environmental, social, and economic. But of course, there are a lot of conflicts of interest between those three dimensions and it’s naive to expect simple ‘win-win-win’ situations. moreover, you can criticize this definition as ‘anthropocentric’ — it centers on current and future generations of *humans*; the well-being of the environment has no intrinsic value. Emphasizing that human civilization depends on the environment is — strictly speaking — anthropocentric. I am not against anthropocentrism as such, but we just need to be clear that we only assign the environment an instrumental value when arguing like that.
MD: Now that we have the definition out of the way, can you please share what does sustainability mean for technology, specifically AI?
DB: It means that AI must be developed and used in a way that is in line with the requirements of sustainability. AI must contribute to the satisfaction of needs along three dimensions, today and in the future. However, this is a simplified statement because every #technology involves trade-offs. There is not one tech that satisfies everybody’s needs, including the environment’s, now and in the future. But it’s important to add the sustainability dimension to AI because AI is mostly presented as a solution to sustainability issues…
MD: Agreed. No technology solution is perfect, and neither is AI, which also has certain tradeoffs. How concerned should we be about AI’s carbon footprint?
DB: We should be very concerned, especially when we waste tons of CO2 for developing what I’d call “AI for Nonsense”, i.e. AI that’s not satisfying any relevant need, neither now nor in the future & that’s at best a marketing ploy.
MD: Good point. Organizations like the United Nations have also established Global SDGs, how can AI help with these sustainability goals?
In my opinion, environmental sustainability is one of the most promising domains to deploy ‘AI for Good’. the environment is an excellent use case for collecting and analyzing data that help us to better understand and address key environmental challenges. In contrast to the use of AI in ‘human settings’, you typically don’t run into problems of privacy and discrimination when using it for environmental purposes.
MD: How should we (companies, individuals) manage the environmental impact of AI?
DB: They should be transparent about it. As sustainability reporting increases and even becomes mandatory in some countries, it would be desirable to see companies report not just on the emissions of their buildings, factories, etc.. but also of the software, they use. This type of data might not be easy to come by at the moment, but increasing the awareness and addressing the challenge upfront are important first steps.
MD: Thank you so much for that insightful discussion, Dr.Baur. Being intentional about transparency and increasing awareness are great first steps towards a more sustainable future with AI.
Join us again for our next monthly Twitter Chat on Friday, August 30 at 8a PT/11a ET/3p GMT with our expert guest Maria Axente, Program Lead — AI for Good at PWC UK when we discuss how to build AI programs with the greatest impact on social good.
Mia Dand is a digital transformation leader with unique expertise in leading complex cross-functional programs at the critical intersection of business, data, technology, and governance + ethics. As the CEO of Lighthouse3, strategic research and advisory firm based in Oakland, California, Mia excels at guiding global companies on the responsible adoption of new & emerging technologies like AI for successful business outcomes.
Mia is a passionate diversity in tech advocate with extensive experience in building human-centric programs at global companies like Google, HP, eBay, Symantec, and others. She is the creator of “100 Brilliant Women in AI Ethics”, a definitive guide for diversity and soon to be launched “Diversity in AI” online tool to help conference organizers and companies recruit more talented women in this space.
