Advanced Computing in the Age of AI | Sunday, May 26, 2024

AI Saves the Planet this Earth Day 

Earth Day was originally conceived as a day of reflection. Our planet’s life-sustaining properties are unlike any other celestial body that we’ve observed, and this day of contemplation is meant to provide all of us with a reminder that it is our solemn duty to protect all that Earth has provided for us.

While there is much work to be done in terms of shifting toward renewable sources of energy and more sustainable lifestyles for the nearly 8 billion people who inhabit this blue marble, certain organizations are using modern AI tools to take small steps toward building a better world. To reward this noble work, and to provide hope for a better tomorrow, we’ve compiled some of the more interesting stories concerning how AI is being used to save the planet.

IBM

IBM has a track record for AI and data excellence, and it therefore makes sense for the company to put this expertise to good use in the fight against climate change. IBM is using its computing resources for a variety of environmentally-focused jobs.

To begin, the company has the IBM Global Real Estate venture, which tracks, analyzes, and reports progress toward specific sustainability goals. This AI program collects data from more than 6,500 utility bills they receive every year from around the world to summarize total energy consumption, cost, and renewable energy purchases. This allows the company to generate reports and filter this information by location, geography, and utility to allow IBM to understand where energy conservation is most needed.

Additionally, IBM is working with organizations to both accelerate the transition to low-carbon energy sources as well as change current operations to ensure these organizations are more climate resilient. IBM specifically mentions its work with the supermarket chain Salling Group to help balance their electricity consumption in relation to the supply of renewable power sources. Using AI and Internet-of-Things sensors, IBM is providing the ability to make intelligent energy decisions. Large buildings, such as grocery stores, can partially pause their energy use for heating or cooling up to a certain threshold that is determined to not impact operations.

IBM is also working hard to help organizations become more climate resilient by giving them the tools they need to predict the impacts of climate change. The company is working on a geospatial foundation model that will be capable of tracking deforestation, detecting greenhouse gases, or even predicting crop yields. As climate change makes our world more unpredictable than ever before, these AI-driven tools will become ever more important.

NASA

IBM’s work with foundation models in the area of environmentalism has far-reaching implications. Specifically, NASA is celebrating a collaboration with the company to create an AI geospatial foundation model. This model is trained on the large amount of data generated by NASA’s Harmonized Landsat and Sentinel-2, and it provides a base for many AI-powered studies that are focused on environmental concerns.

This model has already achieved a level of success. Since teaming up with IBM in 2023, NASA has been able to use the model to detect burn scars, delineate flood water, and classify crop and other land use categories.

Of course, NASA is overflowing with a litany of data concerning many aspects of Earth’s environmental problems. However, the science organization required large amounts of computing power and AI algorithm optimization to achieve its goals. Thus, the partnership with IBM to provide these much-needed resources was a perfect match.

Google

Humanity owes its existence to historically consistent climate and environmental patterns. While anomalies did arise, much of human civilization has existed within a predictable world. Sadly, even the most modest predictions of climate change’s future come with a level of uncertainty that we as a species largely haven’t experienced in such a short period of time. This unpredictability is manifested most obviously in the increase in extreme weather events.

Thankfully, Google is working hard to help prepare humanity for changes in our weather system. Specifically, the search engine giant has developed the Scalable Ensemble Envelope Diffusion Sampler (SEEDS) model. Similar to large language models like ChatGPT, SEEDS uses generative AI techniques to generate multiple likely weather scenarios at a fraction of the cost of the traditional physics-based forecasting models that are currently used.

Unlike these models, SEEDS is based on denoising diffusion probabilistic models, which are state-of-the-art generative AI methods created in part by Google Research. Conditioned on as few as one or two forecasts from an operational numerical weather prediction system, SEEDS can generate a large ensemble that yields plausible real-life weather scenarios. SEEDS can realistically capture the geographical covariance and the correlation between mean sea level pressure and mid-tropospheric geopotential by directly modeling the joint distribution of atmospheric states. Weather forecasters typically utilize both of these closely connected criteria to assess and validate their forecasts.

The climate crisis poses one of the largest obstacles to continued human existence that our species has ever faced. However, by listening to our scientists and using our most valuable and modern tools, we can begin to build a sustainable world for generations to come. On this Earth Day, let us take some time to reflect on what is most important to humanity as a whole. These companies prove that there is hope for us yet if we work together and strive toward a better future.

EnterpriseAI