AI: A Potential Game-Changer for Climate and Environment
Artificial intelligence (AI) has emerged as a pivotal technology in both exacerbating and potentially mitigating environmental issues, according to a comprehensive survey conducted by Forrester. The report, titled "How AI Will Accelerate The Green Market Revolution," delves into the dual impact of AI advancements on planetary health and offers strategic guidance for tech and business leaders on adopting 'green AI' to minimise environmental damage.
The Environmental Cost of AI
In recent years, the environmental toll of AI has skyrocketed. According to Forrester, companies like Microsoft and Google have witnessed substantial increases in their greenhouse gas (GHG) emissions, attributed primarily to the construction of data centres necessary to support advanced AI and cloud computing systems. Specifically, Microsoft saw nearly a 30% rise from 2020 to 2023, while Google's emissions surged by 50% over the past five years.
The International Energy Agency predicts that global electricity demand for data centres, driven by AI's growth, will double between 2022 and 2026. To put that into perspective, this consumption will be comparable to the current electricity usage of Germany.
Moreover, the increasing demand for AI could severely impact water resources. Water-based cooling systems, although efficient, are poised to consume between 4.2 billion to 6.6 billion cubic meters of water by 2027, equivalent to nearly half the UK's annual consumption. This growing water demand is already causing friction in arid regions such as Arizona in the United States and Spain.
The production of AI-related hardware—including semiconductors, servers, and smartphones—also contributes significantly to environmental damage, even more so than data processing activities. AI finds utility in various industries, including fossil fuels where it aids more efficient extraction methods, paradoxically exerting both positive and negative impacts.
The Positive Potential of AI
Despite these challenges, AI holds considerable promise in advancing environmental sustainability. Forrester highlights the potential for AI to contribute to reducing GHG emissions by 5% to 10% by the year 2030. The technology can play a critical role in climate-related adaptation and resilience initiatives, including assessing climate risks, preventing deforestation, predicting wildfires, and providing early flood warnings.
Machine Learning (ML) models, for instance, can integrate weather data with low-orbit satellite imagery and IoT sensor data to enhance the prediction accuracy of natural disasters such as floods and wildfires. This allows local authorities to take pre-emptive actions and optimally allocate resources.
Computer vision, powered by ML, is effective in direct sustainability efforts like analysing satellite images to gauge forest conditions and indirect initiatives such as minimising waste in manufacturing processes. Additionally, ML aids in more accurate consumer demand forecasting, thereby optimising supply chains and reducing carbon emissions.
Deep Learning (DL) also offers transformative possibilities. By analysing vast amounts of historical and real-time data, DL can predict severe weather events with greater accuracy, reducing climate and physical risks for critical assets.
Strategies for Embracing Green AI
For organisations looking to mitigate the environmental impacts of AI, Forrester suggests a multi-faceted approach. Companies should evaluate the ecological footprint of their AI activities using tools such as GenAI Impact's EcoLogits calculator. Transparency and accountability can be enhanced through the implementation of sustainability management software and cloud reporting tools.
The report advises against hastily integrating AI technologies without considering their environmental impact. Instead, businesses should select AI models judiciously, ensuring they are appropriate for specific use cases to minimise the use of servers, clouds, and devices.
Prioritising the use of efficient hardware and opting for low-carbon regions are also crucial for fostering green AI adoption. Organisations are encouraged to focus on applications compatible with older hardware to mitigate the environmental impact associated with constant hardware upgrades.
In conclusion, while AI presents considerable environmental challenges, it simultaneously holds the potential for significant positive impacts on global sustainability. Careful and strategic implementation of 'green AI' practices can enable organisations to harness AI's power responsibly.
Industry stakeholders and marketers are invited to explore these trends and strategies further at Digital Marketing Asia 2024 in Hong Kong, scheduled for 22-23 October. This event promises valuable insights into sustainable growth in the digital era, with opportunities to network and share innovative ideas with industry leaders.
Source: Noah Wire Services