AI accelerates the development of sustainability solutions

- Did you know that by 2030, the world will need over 90 times more battery capacity, a 25% reduction in the carbon intensity of steel, and 30 times more durable carbon removal capacity compared to where we were in 2020? These are just some of the profound challenges that can only be overcome with new sustainable materials, processes, and products. However, as I explained in the previous video, there's a problem. Current approaches for developing and deploying sustainability solutions can be painfully slow and costly. Consider, for example, the development of new materials such as those that are low in carbon or ones that can absorb carbon. IBM research has estimated that on average it takes roughly 10 years and 10 million to a hundred million dollars to discover just one new material with specific useful properties. Given the pace at which the world needs to transition to sustainability, this might seem discouraging, except AI tools are already making these types of discoveries faster and cheaper. They're helping researchers to collect and analyze massive data sets, generate and test hypotheses, design, optimize, and automate experiments, and discover new insights and theories. Let me give you two examples to show just how powerful these new approaches to scientific discovery can be for accelerating the deployment of sustainability solutions. First, consider the world's food systems, which are facing increasing risks from climate change, increasing demand from population growth, and demographic shifts. Meanwhile, the global agricultural food systems account for nearly a third of all greenhouse gas emissions. This poses an enormous challenge. By 2050, under less favorable climate conditions, global food production needs to be 50% greater than it is today. While at the same time emissions reductions need to be cut to near zero, new varieties of crops will be a key part of meeting this challenge. But traditional breeding methods for improving crops are slow, usually taking seven to 12 years to create new varieties. AI is being used to speed this process up. It is helping researchers discover more climate resilient crops by analyzing large, complex data sets, finding genetic markers linked to beneficial traits, and simulating how plants will react to environmental stressors. Batteries provide another example. Batteries are essential for storing and delivering energy for solar, wind, and other sources of renewable energy. However, today's batteries require scarce and expensive materials, such as lithium. AI is being used to accelerate the discovery of new materials for batteries by enabling researchers to search through millions of potential candidates, predict their properties and performance, and identify the most promising ones for further testing and development. For example, researchers at the Pacific Northwest National Labs in partnership with Microsoft used AI tools to narrow down millions of potential new battery materials to a few handfuls of promising candidates in less than a week. A screening process that could have taken more than 20 years to carry out using traditional research methods. These examples only scratch the surface. From helping to find new catalysts that enable low cost production of green hydrogen and low carbon industrial materials, or even hasten the arrival of nuclear fusion. The possibilities are endless. Given the scale and pace of change that the world needs, AI's potential to transform scientific discovery is a game changer.