The UN’s 29th annual Climate Change Conference of the Parties (COP) was not without controversy. Bringing together climate scientists, business leaders, Indigenous representatives, and activists, the 2024 conference highlighted the need for real action against the climate crisis and the importance of using technology that helps our planet and its people. 

Oil: God’s gift? 

Climate leaders at COP28 in 2023 called for the tech industry to take action against the climate crisis in light of their role in the long-term shift in average temperatures and weather conditions.

2024’s COP29 caused much upset which dominated the conference’s conversation. President of Azerbaijan and host of COP29 Ilham Aliyev praised oil as a “gift of god” and prompted similar statements from the Organization of the Petroleum Exporting Countries’ Secretary General Haitham al-Ghais and the Natural Gas lobby unsurprisingly in favour of fossil fuels. Coupled with underwhelming climate finance targets, these comments undermined the conference’s ambitious proposal for reform in the tech industries. The Alliance of Small Island States (AOSIS) walked out after being denied increased funding to fight the climate crisis. When they returned, the conference agreed on $300 billion of funding, when $500 billion was called for.

Technology has the potential to both accelerate and hinder climate progress. Implementing technological solutions is essential to curbing the effects of the climate crisis, a discussion that Digitalisation Day aimed to provoke. 

The establishment of Green Digital Action in 2023 prompted the first-ever Digitalisation Day 2024 at COP, outlining ways digital technologies can boost climate action. Notably, the COP29 Declaration on Green Digital Action received over 1,000 endorsements from governments, companies, civil society organizations, international and regional organizations, and other stakeholders. 

But what is a declaration or an endorsement worth? 

Empty words

Ultimately, declarations from the UN are not legally binding. An endorsement doesn’t require a country to incorporate a document into their nation’s laws. I feel that an endorsement serves as a glorified pat on the back for world leaders, leaving them with a false sense of accomplishment. 

I feel that the COP29 declaration is littered with a slew of pragmatic phrases that, while highlighting the environmental costs of artificial intelligence (AI), fail to provide a framework for implementing tech-based solutions for the climate crisis. More importantly, endorsing a COP declaration doesn’t hold world leaders accountable for reducing greenhouse gas emissions. 

Many in the environmental activism space have grown disillusioned with COP conferences and the ambitious goals they set out. As such, many have begun to adopt the attitude of climate activist Greta Thunberg, who encouraged the public to ‘stop pretending’ that the COP process will lead to meaningful action on the climate crisis. Our climate leaders continuously warn us to halve greenhouse gas emissions by 2030 to avoid climate catastrophe. And amid such warnings, many feel there is little time to waste.

Can green AI be our alternative future?

However, there exists an alternative future. One where rising electricity needs incentivize us to create green , based on renewable energy sources, like wind and solar, and increase the efficiency of AI, so that it uses less power and energy. Even as AI expands its reach, we can dramatically reduce the amount of greenhouse gases it creates. This concept isn’t a novel one. The need for green AI was also a key topic during the AI for Good Global Summit, held by the International Telecommunication Union — the UN’s agency for information and communications — in cooperation with other UN agencies earlier this year.

One initiative — a collaboration between Google and the World Meteorological Organization — predicted riverine floods up to seven days in advance by combining two AI models: the Hydrogenic Model and the Inundation Model. 

Using the Hydrologic Model, Flood Hub can foresee the amount of water flowing in a river. With the Inundation Model, the AI shows what areas will be affected and how high the water level will be. Crucially, the model provides warning before climate disasters occur, limiting deaths and damage in the 80 countries where FloodHub is available. Flood warning systems will become increasingly essential in the coming years as the climate crisis ramps up the frequency and intensity of weather events. 

Another Google operation called Project Contrails was created to limit the environmental impact of airplane usage. Condensation trails — or contrails — are clouds formed behind airplanes when water vapour in the air condenses around tiny particles of soot and other pollutants emitted by airplane engines, trapping heat in Earth’s atmosphere. 

Project Contrails uses AI to project when and where contrails will form based on weather data, satellite data, and flight data. It is important to note that contrails only comprise roughly 35 per cent of aviation’s global warming impact, with carbon dioxide emissions being the largest contributor.

AI’s quiet environmental costs 

With AI, power comes at a great cost: energy. Data centres — which house the infrastructure needed for AI computations — used an estimated 460 terawatt-hours (TWh) in 2022 to meet AI demands. In two years, the electricity consumed by data centres, AI, and the cryptocurrency sector will reach roughly 1,000 TWh. This technology offers an undeniable lifeline for predicting and preventing disaster. Simultaneously, building more fossil fuel plants to generate the energy required for AI and data centres further exacerbates the climate crisis. 

Can emerging AI-based technologies aimed at reducing greenhouse gas emissions and rising temperatures create a positive difference? Or are the growing energy and resource demands of data centers too much? Will they be counterproductive to the same climate conditions these models are trying to solve? 

The best time for world leaders to step up and invest in green AI has passed. The second best time? Now.