Artificial Intelligence, Climate Messaging and CARICOM

Our region knows all too well the realities of the climate crisis. Together, as small island developing states, we routinely stand before global bodies, such as the United Nations, to speak about how a changing climate threatens our very survival. The region's primary intergovernmental organisation, the Caribbean Community (CARICOM), has been central to those advocacy efforts, but recent activity of that organisation is contracting the organisation's climate conscious narrative.

Accross the region governmental and intergovernmental agencies are increasingly turning to artificial intelligence tools to generate social media graphics, promotional videos and public awareness flyers. CARICOM appears to have joined the bandwagon recently, evident by content shared to their Facebook page. While the increased speed and cost effectiveness of using AI tools for these purposes are well noted, the environmental cost ( from where I stand at least) directly collides with the sustainability agenda championed by CARICOM.

In a word, I call it "hypocrisy". Here's why.

AI is often percieved as a soley virtual technology. This is far from reality, however, so let's touch the grass while we still have it.

When you ask your AI tool of choice to generate an essay, poster or video, you are engaging with a large scale data centre running specialised processors. These processors perform billions, or trillions, of calculations to fulfill even the simplest requests. This process of computation consumes not only significant amounts of electricity, often fueled by fossil fuel dominated power grids, but also enormous amounts of water for thermal management.

Researchers have found that generating images and videos, the exact kind of uses we are currently seeing from CARICOM, is among the most energy intenstive applications of AI. (O’Donnell & Crownhart, 2025).

But of course, "the Caribbean is only responsible for less than 1% of global greenhouse gas emissions" (tell that to the hurricanes!). Yes, it is true that the region is disproportionately affected by the climate crisis.

And yes, AI data centres are located outside of the Caribbean region. However, the greenhouse gasses generated because apparently CARICOM could not locate a single competent graphic designer, or a Canva subscription, within its member states , enter the same global atmosphere as the bush fire down the road. Carbon emissions in Nevada or Port of Spain tells the same story for the Caribbean region.

Because climate change does not recognize geographic outsourcing.

So when CARICOM uses an AI generated flyer to tell us that it is "enhancing its sustainable energy agenda", we return to my word of the day. Hypocrisy.

How does this align with CARICOM's commitments to the sustainable development goals? In my estimation, it clashes with at least three of them. Here's how.

SDG 6 - Clean water and Sanitation

Millions of litres of potable water are used for colling AI data centres, while Caribbean states face droughts and saltwater intrusion to fresh water. CARICOM advocates for water conservation while outsourcing components of its communications strategy to water intensive foreign infrastructure.

SDG 8 - Decent Work and Economic Growth

The use of AI to generate organisational communications displaces knowledge-based jobs in the Caribbean region by taking work away from illustrators and graphic designers.

SDG 13 - Climate Action

CARICOM leaders regularly call for aggressive global emissions reductions. However, the use of AI generated media depends on high energy data centres powered by fossil fuels. Using these tools for routine communications contradicts the principle of reducing avoidable emissions.

Ultimately, none of this means that CARICOM is responsible for the environmental impact of AI.

However, it raises eyebrows when the organisation says fossil fuel dependence is catastrophic while simultaneously turning to AI tools for non-essential communication outputs.

The real issue is not the adoptation of the technology but proportionality. Very few would object that using AI for hurricane forecasting or disaster preparedness are significant applications that could save lives.

Using AI to produce a flyer for a conference or a message on sustainability would garner far more objections.