How AI will impact our planet and how we can embrace green AI

AI has the potential to unlock insights that could help reduce GHG emissions by 5% to 10% by 2030 and significantly bolster climate-related adaptation and resilience initiatives, said a Forrester survey. 

Conducted by Forrester, the report named "How AI Will Accelerate The Green Market Revolution" reveals the significant impact AI is having on the environment, how this will evolve, and how tech and business leaders can embrace green AI to reduce the environmental impact of their AI initiatives.

AI in general and Gen AI in particular are already having a significant impact on planetary boundaries. According to the report, Microsoft’s greenhouse gas (GHG) emissions jumped by nearly 30% from 2020 to 2023; Google’s have increased by 50% over the past five years. This is mainly due to the construction of data centres that host AI and cloud computing systems. 

Looking ahead, Forrester predicted that Gen AI democratisation will explode energy consumption. The International Energy Agency anticipates that global electricity demand for data centers, driven by AI growth, will double between 2022 and 2026, at which point it will roughly equal the electricity consumption of a country such as Germany.

Furthermore, AI will drain water resources, said the report. Water-based cooling remains the most energy-efficient option for data centres, and its overall impact on water consumption has been relatively low. Until now: Experts predict that, by 2027, surging demand for AI could lead to the withdrawal of 4.2 billion to 6.6 billion cubic meters of water — nearly half of the UK’s annual consumption. The growing thirst of AI and data centres are already causing significant tensions in arid regions like Arizona and Spain, and these can only rise.

Interestingly, producing the semiconductors, servers, laptops, and smartphones that run genAI has a greater environmental impact than data processing. AI will be a double-edged sword in the oil and gas industry as fossil fuel companies can use AI to extract oil more efficiently with less waste, said the report. 

Potentials of AI on our planet

On the bright side, AI has the potential to unlock insights that could help reduce GHG emissions by 5% to 10% by 2030 and significantly bolster climate-related adaptation and resilience initiatives.

According to the report, AI can help assess climate risks, limit deforestation in the Amazon, predict and combat wildfires, and provide early flood warnings; it can also shape the climate-related adaptation and resilience strategies of cities such as Lagos, Nigeria.

For example, ML models can couple weather data with low-orbit satellite imagery and IoT sensor data to better predict floods and wildfires, helping local authorities to take preventive measures and optimise resource allocation.

Powered primarily by ML, computer vision is a key technology for direct sustainability use cases such as analysing satellite images to assess the state of a forest and indirect ones including reducing waste during manufacturing processes. ML also helps to better predict consumer demand and optimise the supply chain, which helps reduce waste and carbon emissions, said Forrester. 

On the other hand, deep learning (DL) uses neural networks to generate hierarchical insights from images, video, text, or audio and thus optimise cradle-to-cradle approaches to monitor complex supply chains in real time or identify more sustainable materials. DL can analyse vast amounts of historical and real-time meteorological and satellite data to more accurately predict hurricanes, tornadoes, typhoons, and wildfires and reduce climate and physical risks for strained assets.

Embracing green AI

In fact, global firms can lower their costs and the environmental impact of their AI usage by embracing green AI. Forrester suggested that companies should measure the environmental impact of AI activities via tools such as GenAI Impact’s EcoLogits calculator.

Companies should also establish transparency and accountability with a cloud sustainability goal. By implementing sustainability management software and cloud reporting tools, businesses give transparency to their cloud sustainability journey.

Marketers are also warned to not jump on the AI bandwagon too quickly. "Don’t use a genAI sledgehammer to crack a simple data nut; select the right model for each use case and prepare to use a family of different models. Tech teams can help business leaders select relevant data, analytics, and AI techniques to minimise the use of servers, clouds, and devices," said the report.

Organisations should focus on efficient hardware and low-carbon regions to foster green AI adoption. Given devices have the most negative impact on the environment, companies should prioritise apps that will work tomorrow with yesterday’s laptops instead of rushing to the latest AI PCs and smartphones, said Forrester. 

Explore the latest digital marketing trends to empower your brand for sustainable growth. Join 200+ industry players and marketers at Digital Marketing Asia 2024 Hong Kong on 22-23 October and discover the key to sustainable growth in the new digital era, network with industry leaders and find out more about real-life marketing wins and thought-provoking ideas.

Related articles:

IMDA pushes for AI adoption and retail transformation
Anthropic launches business plan and app to increase corporate AI adoption