As you’ve probably read, there is an elephant in the room related to the growing integration of AI systems with our economy. Namely, that some of these systems have a large – and often opaque – environmental footprint. There is an ongoing tug of war between the rapidly expanding demand for compute to run AI systems and the associated surge in electricity and water use from the data centers powering them.

But there is another side to the story. While large language models (LLMs) like ChatGPT etc. are quite resource intensive, many smaller, targeted applications linked to sustainable AI are not. Small and medium-sized enterprises (SMEs) can improve team productivity, reduce energy expenses, optimize product development & supply chains, and streamline reporting and stakeholder communication by thinking strategically about sustainable AI. Put differently: many use cases for AI promote resiliency and offer clear sustainability benefits.

My work in this space shows that for most SMEs, AI is top of mind but their leaders are still in the learning phase, undecided on when and where to explore deeper integration. A recent survey confirmed this assessment: the majority of respondents indicated that they had limited integration of AI today but that it was a priority to find ways it can advance sustainable practices. As a consultant, my role is to help companies and organizations bridge this gap.

Every business or organization is different. There is no “one size fits all” AI solution for everyone. However, there are clear areas of operations that can realize efficiency or resiliency gains from AI integration. I call these “AI touchpoints” and here are some examples:

• Energy management systems that predict and optimize consumption patterns. This includes smart building solutions that adjust heating and cooling based on occupancy. Production line optimization to reduce waste and material usage.
• Sustainable supply chains through route optimization for transportation and reduced fuel consumption. Predictive maintenance to extend equipment life and reduce replacement waste. Demand forecasting to minimize overproduction and inventory waste, and to guard against disruptions from supply chain failure.
• Sustainable product development through material science applications to identify eco-friendly alternatives and design optimization for reduced environmental impact. Circular economy modeling to identify reuse/recycling opportunities.
• Environmental Monitoring through satellite imagery analysis to track deforestation, emissions, or water usage. Sensor networks for real-time environmental impact monitoring. Using AI to streamline sustainability reporting requirements and to plan for climate-related environmental impacts to business processes.
• Waste Reduction including improved recycling sorting and processing. Predictive modeling to minimize food waste in retail and hospitality. Process optimization to reduce industrial byproducts and waste
• Carbon Management Emissions tracking and reporting automation. Carbon footprint calculation and reduction recommendation engines.
• Stakeholder Engagement through personalized sustainability communications based on customer preferences. ESG reporting automation and visualization. Employee engagement tools for sustainability initiatives

It is understandable to be hesitant or even overwhelmed by the prospect of AI integration with (or in) the work of your company or organization. Besides the environmental impact of LLMs, AI systems can contain harmful biases, displace workers, and add uncertainty to the structures and processes of how teams work.

All of these concerns are fully justified. Yet recent history shows that failure to adapt to technological change (or embracing it too late) can erode competitive advantage and inhibit innovation. Think of companies or organizations that were slow to embrace a digital transformation in the 2000s and 2010s. Was this caution rewarded with a smarter long-term digital strategy? Or did it slow growth relative to more intrepid peers?

The key to implementing AI is the same as any other new technology: step back, learn about what is available, map out possible “touchpoints” in collaboration with team leaders, and develop a blueprint for action. Young workers today want the companies or organization they work for to embrace sustainability. Smart integration of sustainable AI can demonstrate this commitment, while freeing up time for more value-adding tasks and upskilling.

The energy and resource intensity of AI-trained LLMs is no secret. Training one AI model can consume as much electricity as 100 U.S. homes do in an entire year. Big tech companies own a majority of the data centers that host cloud-based computing operations running these LLMs. Recently launched initiatives like Hugging Face’s AI Energy Score, Google’s 4M approach, and IBM’s Cloud Carbon Calculator assist in measuring and optimizing the energy and resource efficiency of AI systems.

SMEs must consider the extent to which AI integration will impact their scope 3 emissions. AI systems integration with large environmental impacts need to be evaluated against broader sustainability goals. At the same time, there can be costs (some environmental) to not integrating AI: energy consumption isn’t optimized leading to higher utility bills; supply chains are inefficient or produce unnecessary waste; stakeholder communications aren’t automated, leaving employees less time for other tasks.

Finding the right areas to integrate AI with a clear understanding of the associated environmental costs and benefits is a critical task for SMEs in 2025.

The AI regulatory and ethics context is complex and constantly changing. SMEs should stay up to date on the latest developments relevant to their operations.

National governments are proving more aware, yet inconsistent, in their approach to AI regulation when compared to social media a decade ago. In February 2024, the US Congress introduced the Artificial Intelligence Environmental Impacts Act of 2024. The legislation would direct the National Institute of Standards and Technology (NIST) to develop standards to measure and report the full range of artificial intelligence’s (AI) environmental impacts, as well as create a voluntary framework for AI developers to report environmental impacts. But the new Congress shows no signs of advancing this legislation.

The European Union member states approved an “AI Act” that would require “high risk” (which include the powerful “foundation models” that power ChatGPT and similar A.I.s) developers to report their energy consumption, resource use, and other impacts throughout their systems’ lifecycle. India and China have yet to introduce regulations targeted at reporting and managing these AI environmental impacts. Yet AI governance is a clear policy priority in both countries – evidenced by China’s rolling out of binding AI regulations and India’s intervention in the introduction of new AI products from national technology companies – suggesting that national governments are developing strong regulatory frameworks where supplementary environmental standards could be added in the near future.

AI ethical standards are also evolving within national governments, and can be developed on the enterprise or organizational-level too. These standards ensure that any AI systems are fair (without bias), transparent (easy for non-technologists to understand), accountable (with clear responsibilities for different departments or team members), respecting privacy, and aware of the societal and environmental impacts of AI deployment. AI Impact Assessment Tools (like this one developed by Algorithm Watch) are a good place to start.

Contact Jeremy Tamanini to continue the conversation on this topic, as well as reading these related insights from the practice:

AI Everywhere: Tangible Applications for Sustainability Teams (link here)

The AI Elephant in the Room (link here)

AI x Sustainability in Trump 2.0 (link here)

Remarks to the National Sustainability Society (link here)

AI in Building & Construction: Tangible Applications for Sustainability Teams (link here)

How to Work with Satellite-Based Sustainability Data (link here)