We are currently witnessing an unparalleled acceleration in technological advancement. This year saw Bill Gates hail the dawning of the AI era, marking what many believe to be the fourth industrial revolution.
Every day, AI reshapes how we address multifaceted issues in numerous sectors, including medicine, urban development, and environmental science. Particularly in the environmental arena, AI has been instrumental in enhancing climate models, streamlining power grids, pioneering smart buildings, fostering precision agriculture, and employing satellite imagery to monitor our oceans and forests. Its unmatched capacity to unearth patterns in expansive datasets is central to AI’s prowess.
Marrying this with progressive hardware and cloud computing’s agility, we’re empowered to process data at unmatched velocities and magnitudes.
In 2020, a survey revealed that AI could assist in accomplishing 79% of the 169 objectives set out by the United Nations in its Sustainable Development Agenda. These goals are vital for both environmental and economic aspects, and such technologies can be instrumental in addressing climate change and advancing sustainable growth globally.
Is it a Double-Edged Sword?
Every silver lining has a cloud. The technology we harness to counteract environmental concerns may also exacerbate them. For instance, creating and utilizing Large Language Models (LLMs) demand substantial computational power, consuming significant energy, hence creating a remarkable environmental footprint as negative externality.
One underreported consequence is AI’s water footprint. Microsoft’s report disclosed a staggering 30% increase in water consumption between 2021 and 2022, correlating this surge directly to their investment in AI. Another study between UC Riverside and UT Arlington suggested that engaging with OpenAI’s ChatGPT for a simple 20-50 questions conversation might consume water equivalent to half a liter bottle.
To truly leverage AI’s potential for sustainable growth, we need to transition towards ‘Responsible AI’ or ‘Sustainable AI.’ Implementing principles of digital sobriety and setting water and carbon budgets early on are crucial to this endeavor.
We must recognize the environmental impact of these technologies and adopt a responsible approach to their use.
How AI and Sustainability work together
Despite these challenges, we remain optimistic about AI’s role in forging a sustainable future. At Globant, we’ve always believed in the power of sustainable AI. As early as 2019, our AI Manifesto underscored the importance of adopting an eco-friendly approach.
Internally, we have implemented training programs, such as Green IT Training, to raise awareness among our Globers about concepts such as Digital Sobriety and encourage the development of digital solutions with a low carbon footprint.
In addition, we collaborate with our clients and academia to promote best practices. To cite one example, in partnership with researchers at Georgetown University, we are exploring how to minimize our AI solutions’ energy consumption and emissions and lead the way in sustainable tech development.
Globant’s sustainability vision
At Globant, we are committed to leveraging disruptive technologies to address challenges relevant to the well-being and prosperity of humanity. That’s why, in 2019, we launched our Sustainable Business studio to focus on sustainability issues.
In collaboration with Inmarsat, a British satellite services company, we explored how satellite technology and AI can support Net Zero initiatives. We quantified carbon emissions in three key sectors: transport and logistics, agriculture and forestry, and energy systems, which account for 60% of global emissions.
This combination of technologies can reduce 1.5 billion tons of carbon annually, equivalent to the lifetime emissions of 50 million cars. Examples of applications include optimization of transportation routes, energy efficiency, and forest fire prevention.
If these technologies gain widespread acceptance, the possibilities are immense. They could reduce about 5.5 billion tons of carbon annually, a significant one-sixth of the goal to restrict the global temperature increase to under 1 degree Celsius by 2030.
If we look even further ahead, we find additional opportunities to achieve even more significant reductions (8.8 billion tons, the equivalent of nearly a quarter of 2021 global emissions) by applying satellite and AI technology to other use cases, such as:
- Maritime navigation: fuel savings through autonomous navigation ships, estimated reduction of 400 million tons.
- Energy: optimization of energy generation and consumption, reduction estimated at 1.3 billion tons
- Aviation: adoption of the European Space Agency’s Iris technology, estimated reduction of 100 million tons