AI for Good – Microsoft AI for Earth

Microsoft’s “AI for Earth” initiative uses AI, machine learning (ML), and modern cloud technologies to develop solutions for science, agriculture, and species conservation.

---

Artificial intelligence (AI) can have devastating consequences for people and nature, as illustrated in the article “Green AI? What the True Cost of AI Means for Sustainability.” But AI can just as easily help advance climate protection and nature conservation. This article focuses on true Green AI—that is genuinely sustainable AI.

Humanity and the planet face seemingly insolvable challenges:

Biodiversity loss, ocean pollution, the global climate crisis, and a growing world population that is pushing the Earth’s resources to their limits.

We still know too little about the interconnections within ecosystems, simply because they are incredibly complex. At the same time, however, time is running out.

New Hope for Understanding Complexity Through Artificial Intelligence

Microsoft’s “AI for Earth” initiative harnesses the immense computing power of AI, machine learning (ML), and modern cloud technologies to create solutions for existential environmental problems.

The team is developing sustainable solutions for four critical areas: agriculture, water, biodiversity, and climate change. Through the following examples, AI for Earth is working to ensure that AI can truly be sustainable:

• The Future of Agriculture: To feed the growing global population more sustainably, the “Farm Beats” project uses data-driven approaches. In India and the U.S., sensors help precisely measure soil moisture and nutrients. Since conventional sensors are often prohibitively expensive, drones provide precise imagery instead. Unused frequencies from old television and radio networks transmit the data cost-effectively. This creates a detailed “heat map” that enables farmers to use resources more efficiently and sustainably increase their yields.
• A Revolution in Mapping: Monitoring land areas is essential for environmental protection, but it is extremely time-consuming and expensive. AI can now map the entire United States in just ten minutes using nearly 200 million aerial images. In the past, this would have taken months. Researchers can now put the time they’ve saved to better use, while this AI technology will help identify and manage urbanization, deforestation, and environmental disasters in the future.
• Protecting Endangered Species: Analyzing thousands of satellite images or camera trap photos to study animal populations is a tedious task for researchers. Today, AI takes over the task of reviewing images of, for example, snow leopards in Kyrgyzstan or elephants in the Congo. This allows researchers to focus on actual research work instead of searching for the proverbial zebra on the savanna.

The Bottom Line: The Costs and Benefits of Artificial Intelligence

The measurable positive environmental impacts are significant. By the 2024 fiscal year, Microsoft had permanently protected over 6,400 hectares of land, exceeding its original targets by 30%. The company is building sustainable ecosystems around its data centers. Local employees are restoring the environment by planting native vegetation, stabilizing riverbanks, restoring wetlands, and creating green corridors.

AI-powered monitoring identifies endangered species, enabling a rapid response to ecological crises.

In addition, the “Climate Innovation Fund” is financing 90 water projects worldwide, and in partnership with organizations such as “One Tree Planted”, the company is planting approximately 87,000 trees, with AI calculating the ideal locations for the seedlings’ survival.

But the flip side of the coin cannot be ignored. “AI for Good” always comes at a cost—the massive AI boom has its price:

• Rise in Emissions: Since 2020, Microsoft’s total emissions have risen by about 30% through 2025, driven primarily by the construction of new data centers, which requires large amounts of CO₂-intensive concrete and steel.
• Water consumption: Training large AI models is very water-intensive. Estimates suggest that training a model of the caliber of GPT-4 requires about 700,000 liters of water for cooling and power generation.

Microsoft remains committed to its ambitious goals: By 2030, the company aims to be carbon-negative and “water-positive.” Through its carbon removal efforts to date, the company has already removed 22 million tons of CO₂ from the atmosphere. Under these conditions, artificial intelligence can be sustainable in theoretical terms.

AI is a fascinating paradox of modern society: Enormous negative environmental consequences resulting from massive resource consumption are offset by a significant, sustainable impact. The coming years will show whether the positive effects will outweigh the negative ones in the long term.

Cover photo: CC Hans Veth, Unsplash; edited by the Tina Teucher Team

---

Tina Teucher is a speaker and moderator specializing in sustainable business. In her keynote speeches, she inspires audiences with her insights into the intersection of artificial intelligence, sustainability, and megatrends, and highlights the opportunities presented by an ecological transformation. As a co-founder of startups such as the digital company Future Cooperative (product: Future Maps), Tina Teucher also brings her practical business management experience to the stage.

---