Sustainable business is a strategy that incorporates environmental, social and governance factors into decision-making, and it is becoming an increasingly important component of business strategy. In fact, in a recent Gartner survey, chief executives identified environmental sustainability as a top 10 business priority for the first time in a decade.

Technology is an essential part of the framework that business leaders need to deliver on sustainable business outcomes. However, technology can be a double-edged sword when it comes to sustainability. It can support sustainability goals by improving the quality, scale and impact of environmental initiatives.

For example, organizations can use advanced analytics to monitor enterprise energy consumption and identify opportunities for increased efficiency. However, compute-intensive software or hardware can have a material impact on an organization’s overall carbon footprint.

Artificial intelligence is one such technology that poses both a problem for and solution to climate change. Data and analytics leaders must consider how AI can be used responsibly, sustainability and still generate business value, particularly to support ESG initiatives. Here are the steps that D&A leaders, in partnership with chief information officers and sustainability leaders, can take to make AI environmentally sustainable.

Improve AI model efficiency to reduce carbon output

In a recent paper, researchers performed a life cycle assessment for training several large, common AI models. They found that the process can emit more than 626,000 pounds of carbon dioxide equivalent — nearly five times the lifetime emissions of the average American car.

To support enterprise environmental sustainability goals, it is essential that organizations reduce AI’s carbon outputs. AI models must be as efficient as possible, so that model training does not require large amounts of energy or compute power for marginal increases in accuracy and performance.

Different AI modeling techniques can help to reduce the carbon footprint of an AI workflow. For example, an optimized federated machine learning model could reduce energy consumption costs. Or a transfer learning approach could help redistribute model training costs over several use cases or organizations, reducing the effort required to reinvent model training cycles. Connectionist approaches through composite AI, such as using machine learning with an optimization engine, can also offer energy savings.

When fine-tuning AI models or considering different training techniques, conduct an efficiency-versus-accuracy study to decide if additional resource utilization is justified both from a business and an environmental standpoint.

Streamline compute environments to support sustainability

It can be challenging for many enterprises to adjust their AI models to be more efficient, as most organizations using turnkey solutions will rely on third-party modeling services. However, organizations can control the environment in which they do their modeling.

For example, some vendors have created a high-performance and high-efficiency chip specifically designed for AI modeling. Using custom silicon, organizations can reduce the power usage of a modeling exercise. Hardware that is designed and optimized for machine-learning tasks can offer significantly reduced power consumption over all-purpose processors.

Choose a provider that offers products or services that reduce power consumption during training, such as by compensating on the carbon footprint of AI models by contributing financially to environmental sustainability efforts. Additionally, consider using renewable sources of energy such as solar, hydroelectric or wind to power AI training, infrastructure and compute.

Explore the use of AI to support enterprise ESG goals

Gartner predicts that through 2026, organizations will increase their investments in D&A services by 45% to become more data-driven, digital and ESG-compliant. When used responsibly, AI is among the D&A technologies that can support enterprises in reaching their sustainability goals.

Emerging AI technologies and applications can help improve business operations and optimize processes, reducing carbon and environmental footprints and mitigating material risks. For example, AI-enabled software solutions are emerging that support increasingly complex and comprehensive sustainability reporting, analytics and accounting. AI can also be used in industries like in manufacturing, agriculture, utility or transportation to improve process efficiency and increase automation, reducing carbon outputs.

However, the use of such solutions comes with the risk of negating the energy efficiency gains because of the compute and processing power that may be required. It is essential that organizations streamline AI models and compute environments first, to ensure the responsible use of AI-enabled sustainability solutions.

By relying on AI’s “intelligence” to support efficient power utilization during computation, data movements and power-intensive activities, enterprises can mitigate the environmental concerns surrounding this technology. This adds a new dimension to AI’s business value generation in the form of lesser climate impact. As AI models become more efficient and enterprises establish best practices around AI for the environment, AI can provide additional business value by applying its resourcefulness to support ESG compliance through emerging technologies.

Farhan Choudhary is a principal analyst at Gartner Inc., with a research focus on operationalizing machine learning and AI models, hiring and upskilling, platforming and tooling, and techniques to achieve success with data science, machine learning and AI. He wrote this article for SiliconANGLE. Additional analysis on data and analytics trends, including AI, will be presented during the Gartner Data & Analytics Summit taking place Aug. 22-24 in Orlando, Florida.

Image: PIRO4D/Pixabay