According to GlobeNewswire, the global AI market is projected to grow from USD 273.6 billion in 2023 to USD 5.26 trillion by 2035, reflecting a compound annual growth rate (CAGR) of 30.84%. However, what is the environmental price for such growth? And how can the Caribbean region, as users of AI, contribute to decreasing its environmental footprint?
Artificial Intelligence (AI) is rapidly reshaping our world, promising transformative advancements across industries. However, this technological marvel comes with a significant and growing environmental cost.
Although many of the major AI platforms are managed from developed countries and appear to use the resources where they are based, developing nations also use these platforms, as well as those operating locally. Moreover, for Caribbean and other developing countries that are already vulnerable to the impact of climate change, understanding and managing their contribution to AI’s environmental footprint is not only a responsible course of action but is also crucial for the longer-term sustainable development of our societies and the world at large.
The hidden costs of AI
Generally, the environmental impact of AI stems from three key areas: energy consumption and the associated carbon footprint; water usage; and electronic (e-)waste.
Training and running complex AI platforms, such as large language models (LLMs) and generative AI, requires immense computational power for the vast data centres that underpin their operations.The electricity consumption of data centres globally is projected to continue rising dramatically, with a significant portion of this increase driven by AI.Several reports from the International Energy Agency (IEA) and other organisations predict that global data centre electricity consumption will roughly double by 2030 compared to 2024 levels. If this energy continues to come from fossil fuels, the carbon emissions will be substantial.
Additionally, data centres require large quantities of fresh water for cooling systems to prevent overheating of servers. This water is often evaporated, leading to considerable water consumption. Projections indicate that by 2027, AI-related water consumption could approach that of several European countries (Source: DevelopmentAid).. For countries or regions already facing water scarcity, this demand can strain municipal supplies and have adverse effects on the local ecosystems.
Finally, the rapid evolution of AI necessitates frequent upgrades to hardware, including specialised processing units, leading to an accelerating generation of e-waste.This waste contains toxic materials like lead and chromium, which can leach into the environment, contaminate the food chain and harm human health. Furthermore, such occurrences are more likely to occur in developing countries where ad hoc recycling practices are prevalent. Without proper mitigation, researchers estimate that generative AI alone could create between 1.2 and 5 million tonnes of e-waste between 2020 and 2030 (Source: Scimex).
The unique challenges of AI for the Caribbean and developing countries
As previously mentioned, Caribbean and other developing countries tend not to be the primary hosts for the major AI platforms that we all enjoy, and so it could be argued that the responsibility to manage the environmental fallout lies in the countries where these platforms are built and hosted. However, as users of the technology, and so contributing to the impact on the natural environment, we cannot absolve ourselves of some obligation.
Further, it must be highlighted that most developing regions, and more so the Caribbean, face particular challenges that limit their ability to fully operate large AI platforms and so may, to some degree, always be reliant on the infrastructure and resources in other regions. For example, most Caribbean countries lack the robust and clean energy grids, advanced waste management facilities, and financial resources to implement sustainable AI practices on a large scale. Further, the region is grappling with the existential threats of climate change, in the form of more powerful hurricanes and rising sea levels. Should substantial energy and water resources need to be diverted to power and cool AI systems, critical sectors like agriculture, healthcare, and disaster preparedness could be adversely affected. It is thus incumbent on the region to not only be prudent in its use of AI but also proactive in mitigating the harmful effects on the environment.
Key strategies for more sustainable AI
To manage their contribution to AI’s environmental footprint, Caribbean countries can implement several key strategies, including those outlined below.
1. Prioritise Energy Efficiency and Renewable Sources
Noting the growing interest across the region in increasing its stock of data centres, it is recommended that the erection of green data centres be encouraged, which, among other things, would be more energy-efficient and powered by renewable energy sources like solar, wind, and geothermal. Additionally, localised AI processing, such as using edge computing solutions that process data closer to its source, should be promoted to reduce the need for large, centralised data centres and the associated energy consumption.
2. Implement robust e-waste management
E-waste management has been a growing concern in the Caribbean region, and although some systems are in place, they are not comprehensive and present in all countries. AI is now challenging our countries to up our game by,
- Developing our local recycling infrastructure
- Developing a circular economy ecosystem.
- Raising businesses’ and individuals’ awareness of responsible e-waste disposal and the environmental risks associated with improper handling.
3. Develop context-specific policies and regulations
Finally, as much as we are enjoying the benefits and leveraging the opportunities AI has to offer, in the Caribbean region, we still have not established comprehensive policies or frameworks to guide our use of this emerging but powerful technology, the abuse of which could have far-reaching and long-term consequences. Although the following three recommendations are often made, they bear repeating due to their importance as critical first steps countries ought to take:
- Integrating AI into national development plans
- Establishing ethical AI guidelines
- Actively participating in global dialogues and international discussions on sustainable AI.
In summary, the environmental impact of AI cannot be avoided, and the onus is on all of us to be proactive in doing our part to minimise the impact on the natural world and existing environmental vulnerabilities. For Caribbean countries, the path forward demands clear intentions being established and a conscious and sustained effort being made to ensure that the promise of AI does not come at the planet’s expense.
