AI for Sustainability and the Sustainability of AI

In case you missed it, artificial intelligence is becoming increasingly ubiquitous in our world today. It is reshaping industries how we know them and creating efficiencies never seen before. AI is being used around the world to deliver on use cases for the betterment of society often clubbed under the umbrella of AI for Good.

As AI takes over the world, various new discussions are popping up, from building Responsible AI to how governments should balance innovation vs. regulation. AI is impacting the job market, and how we learn and apply our skills to gain new jobs will also change in the future. Learning new skills requires an entirely different approach than we used to in the last century. AI is creating a new set of winners and losers, and there is discussion on ensuring that no one is left behind and that no single group is at a more significant disadvantage than the other.

EU introduced its first-ever AI Act, which aims to introduce fairness and transparency. As part of the regulation, there is greater emphasis that nobody is denied basic services due to a self-fulfilling loop. There is great focus on responsible AI across the globe.

Sustainability: Another Dimension of Responsible AI

While responsible AI has many dimensions that people talk about like fairness, privacy, transparency, explainability and beyond. Today, we are going to look at another dimension of responsibility, i.e., sustainability. Training a single AI model can consume more electricity than one hundred American homes in one year. Demand for AI services is expected to rise by 30–40% annually over the next 5–10 years. And more powerful AI models will require more energy.

As per World Economic Forum report, the computational power required for sustaining AI's rise is doubling roughly every 100 days. We must balance the progression of AI with the imperatives of sustainability. There are steps we can take immediately — and foundations that we can lay today that will pay dividends in the future — to manage AI and the green transition.

Three Waves of AI Ethics

In the paper Sustainable AI: AI for Sustainability and the Sustainability of AI the authors discuss the three waves of AI ethics:

1. The first wave of AI ethics focused on what AI might do (e.g., Superintelligence) and amounted to ethics. This wave discussed the morality dilemma of AI. What kind of decisions should AI be able to make?
2. The second wave of AI ethics addressed the practical concerns of machine learning (ML) techniques: the black-box algorithm and the problem of explainability. No one should be on the receiving end of an algorithm decision that cannot explain to them why they have received a certain outcome.
3. This third wave has sustainable development at its core. While there is a growing movement to direct AI usage towards ‘good’ uses (i.e. AI4Good) towards sustainable development goals, for example, it is time to move beyond that and address the sustainability of developing and using AI systems in and of themselves. This wave also concerns the climate impact of AI and how to ensure that AI doesn't become an energy sink of the world.

Working with LinkedIn , I am building two new courses on the LinkedIn for Learning platform on

1. Building Sustainable AI: Optimizing the resource consumption of your AI Models
2. Conducting a sustainability audit in your enterprise.

The primary goal of our course is to analyze the energy optimization of AI models. However, the same best practices can be applied to reduce costs. Often, energy consumption and the dollar are directly proportional. While regulations around sustainability are slowly being rolled out worldwide, following the best practices will make you a green company and impact your bottom line.

Always remember, "It is always a good business to be a good business."

Please stay tuned for the course release!

If you like, please subscribe to the FAQ on Data newsletter and/or follow Fawad Qureshi on LinkedIn.