 Hi, my name is Hendrik Hammann. I'm the Chief Science Officer in IBM Research for Climate and Sustainability. I'm here to answer three questions about climate and sustainability and the use of AI. AI can help to address climate change in two major ways. On one hand, we can use AI to accelerate the development of applications which will help with climate and sustainability. But perhaps even more important, AI will help to accelerate the discovery of new solutions to climate and sustainability. Large language models are very different from climate foundation models and the reason is the nature of the data. In large language models, all the context of the information you need to address different downstream tasks is in the sequence of words and you learn the relationship between these words. Now climate, the data is fundamentally multimodal. It's not just text, but it is much more, it is imagery, it is vector information. It comes in all kinds of different dimensions. It can be a sequence, it can be a time series, it can be geospatial. It has different parameters like wind, temperature, precipitation. So it is really complex. Now to unpack all the knowledge in these different data sets, you need to teach how you can actually learn the relationships between all these different dimensions. So can you learn from these large data sets how summer heats in Tokyo are related to winter storms in Chicago? Yeah, first of all, tracking, accounting, measuring greenhouse gas emissions is extremely important because that's where everything starts. Now AI brings three components to this. It brings speed, scale and discovery. So we're using foundation models to accelerate massively carbon accounting. So knowing your carbon footprint from months to minutes and seconds. And that is really important because it enables you to make decisions in seconds. It enables you to make decisions to immediately help to reduce your carbon footprint. Decisions of business processes like business travel, order fulfillment, supply chain, et cetera. Now we're also using foundation models to discover emission sources which are not known. So we use satellite observations to build foundation models which then allow us to pinpoint in space and time and find emission sources to quantify how much emissions are being produced when and where. And that is extremely important but we cannot only do this for the sources but we can also do it for the sinks. So where does carbon gets removed? For example, through vegetation. Now these types of technologies are extremely important for carbon markets which are based on verifiable measurements. And that is a really powerful strategy because now you can put a price on the emissions you produce and a value on the emissions which you remove.