AI's Environmental Impact: How to Cut It

Conversations around AI in business often focus on data security, ethical considerations, and over-reliance risks. However, AI's environmental impact is also a growing concern. It significantly affects energy and water consumption, contributing to global greenhouse gas emissions. Therefore, environmental sustainability must be part of discussions about responsible AI.

There's also optimism that AI can help transition to a lower carbon economy. As AI's influence grows, it's crucial to consider how it both contributes to and combats climate change.

AI's Energy and Water Demands

Increased demand for cloud computing and digital services has driven energy optimization, with data centers adopting energy-efficient hardware and cooling systems. However, AI's energy needs, supported by data centers, are outpacing these efficiency gains.

A single ChatGPT query uses approximately five times more electricity than a web search. Training a language model like GPT-3 uses electricity equivalent to the annual power consumption of 130 US homes. Generating an image requires significantly more energy than generating text. For example, generating text in response to 1,000 prompts requires as much 16% of a full smartphone charge,generating an image requires the equivalent of 100%.

AI operations also require substantial water for cooling data centers. GPT-3 may use about one 16-ounce bottle of water for every 10-50 responses. With billions of queries, AI's projected annual water withdrawal could reach 6.6 billion cubic meters by 2027.

Operational logistics, including construction and maintenance, can account for up to two-thirds of a data center's lifetime emissions. Building infrastructure, manufacturing IT equipment, and chip production also contribute significantly to AI's environmental footprint. As AI workloads increase, so do their power demands.

The World Bank estimates that the broader information and communications technology sector, including AI, generates at least 1.7% of global emissions. While this seems modest, current figures don't reflect future AI consumption. AI adoption is accelerating, along with increased digital penetration, cloud storage, IoT adoption, and cryptocurrencies, potentially amplifying the impact on climate change, even with environmental efficiencies.

The International Energy Agency (IEA) predicts that global data centers may consume up to 1,000 TWh of electricity in 2026, a 400% increase from 2022. Major tech companies like Google and Microsoft have reported year-on-year emissions increases, posing challenges to their climate commitments.

What Can Be Done?

Many businesses are finding new ways to respond to the impacts of their energy use. Organizations associated with cloud computing and AI development are taking strong stances on tackling emissions and climate change. Industry leaders can provide inspiration for better energy management and emissions reduction.

Organizations can take action to address the environmental impact of AI use as part of global sustainability efforts.

How AI Can Help

AI can also be a powerful tool against climate change. It can quickly analyze large datasets to provide insights into climate trends and the effectiveness of emissions reduction strategies.

Additional capabilities for AI to positively contribute to fighting climate change include:

Tracking and reporting energy use and emissions, is best practice for organizations wanting to tackle AI-driven emissions. Publishing an annual emissions inventory or becoming ISO 14001 certified demonstrates a commitment to sustainability. ISO 42001 certification shows a commitment to ethical and trustworthy AI systems. Staying updated on sustainability and AI regulations is also crucial.

AI’s energy and water consumption is a significant issue for ICT corporations and their climate goals. From startups to global enterprises, efforts to reduce resource consumption and emissions driven by AI are essential to preserve our planet. Learning from leading businesses and adopting practical strategies can help align AI strategies with sustainability goals.

The time to act is now.