The technological advancements that humankind has witnessed, from inventing the wheel to testing self-driving cars, make for a story we can all be proud of. Over time and with experience, we seem to have learnt to observe our systems holistically and evolve them in an interconnected manner, understanding that the whole is greater than the sum of its parts.
From a socio-technical perspective, if we chronologically trace the technological revolutions, we will find that the driver behind the events was at first, our basic needs, then our wants, and soon our desires. From basic need of food, shelter, and clothing, to the desire of 10-minute delivery. Every technological upgrade that has been adopted at an unprecedented scale and at a lightning-quick rate by humans has so happened because it made our life simpler, more secure, and more comfortable.
Compared to the 1700s, the standard of living on planet Earth has improved tremendously. The average life expectancy of the world has increased from ~28 years in the 1700s to ~71 years. The chances that a newborn survives childhood have increased from 50% to 96% globally. And with the kind of technological advancements we are foreseeing, life on Earth is going to only get better from here for all sections of society.
However, there is a great paradox to this fairytale. As we enhanced our life on Earth, the Earth itself started getting negatively impacted. Today, despite having all the services at our fingertips, we find ourselves stuck in the midst of such complicated and severe environmental issues that our very existence could soon be under threat and we will be forced to question, all this advancement, for what?
As we approach the next big disruption led by AI, we are facing a familiar path. ChatGPT saw a record-breaking adoption since its launch, reaching over 100 million users within just three months of its launch. Businesses in all sectors are rapidly integrating the tool into their systems with a singular objective - increasing profitability. But recent research revealed that ChatGPT consumes half a litre of fresh water for every 20-50 questions it answers. Imagine the scale of this consumption as GPT’s adoption increases further. For a world whose many regions are enduring the worst droughts they’ve ever faced, this application of artificial intelligence appears to be solving for convenience rather than climate. Overall, AI’s carbon footprint is now larger than the entire airline industry, and a single data centre might consume an amount of electricity equivalent to 50,000 homes.
Can we pause, reflect, and rethink the advancements of AI from a climate-first perspective?
Of course, it’s an extremely idealistic thought, especially in a capitalistic world, that a tool that
can make processes so much more efficient should not be utilised because it is unsustainable. Beyond convenience, AI is revolutionizing healthcare, offering improved diagnostics, personalized treatments, and enhanced patient care. AI-powered tools and robotic systems streamline medical processes, improving efficiency and reducing errors.
AI has so many practical applications to solve a variety of environmental issues. ChatGPT itself can assist organisations in making sustainable decisions. More prominently, there are robust AI methodologies identified by researchers that can accelerate sustainable development. Enabling companies to minimize carbon emissions across their entire value chain, improving extreme weather forecast predictions, providing support tools to the government to respond to natural disasters, and developing plans to mitigate the effects of climate change are just some of the thought starters about AI for the planet. There has already been some progress in this direction. A simple Google search will give a long list of AI-powered climate-focused startups backed by well-intentioned impact investors.
But will there really be an impact if the AI infrastructure itself is unsustainable?
Once again, we find ourselves staring at a fascinating paradox. We have new intelligent tools to save our planet, but these tools themselves can be quite harmful. This is why AI for climate requires a systems thinking approach. Systems thinking is an approach to problem-solving and understanding complex phenomena that focuses on analyzing and understanding systems as a whole, rather than looking at isolated components or individual parts. It involves perceiving the interconnectedness, interdependencies, and feedback loops within a system. It helps us avoid a counter-intuitive event wherein the result is the stark opposite of the objective. Such an event happens due to an isolated approach.
For example, suppose a group of impact investors back an electric vehicle startup to reduce carbon emissions released by ICE engines and contribute to a more sustainable transportation system. However, a counterintuitive event occurs when it is discovered that the increased demand for EVs leads to an increase in mining activities for rare earth minerals used in EV batteries, which emits greenhouse gases to the tune that the net effect compared to ICE emissions barely improves, and there’s hardly any true impact created.
While elements of systems thinking have been applied in certain domains and disciplines, it is fair to say that the world humans have created today has generally not been built with a comprehensive systems thinking approach. Even as we started developing our systems in an interconnected manner, we somehow missed considering the impact on the biggest system of all - our planet.
Organizations across the globe are now resorting to systems thinking as we advance further in AI. The United Nations Coalition for Digital Environmental Sustainability (CODES) highlighted Systems Thinking as a tool in the first of three phases in a shift towards a sustainable planet in the digital age. BCG developed a framework for using AI in combating climate change for their latest AI for the Planet report. This framework is built on a systems-level view of the world and comprises three main themes: mitigation, adaptability and resilience, and fundamentals.
Deployed in a human-centric, responsible and holistic way, AI is an accelerator for sustainable development. Impact investors, global leaders, and entrepreneurs can certainly play a role in developing AI for climate first and convenience second. But they must consider the entire life cycle of AI systems, from the development and training of models to their operational deployment and eventual retirement. By applying systems thinking, stakeholders can identify strategies to optimize energy efficiency, use renewable energy sources, and ensure responsible data centre management to minimize the environmental impact of AI infrastructure.