On March 4, 2025, experts from various fields gathered at Columbia University to explore major questions in the field of inauguration. Columbia AI Summit. From medical, business and policies, science, engineering, and humanities, Summit provided 360 degrees of views on the modified impact on AI’s society.
Climate School Research Institute, afternoon session, From chaos to code: How AI can tame a climate crisisWe mentioned how the AI is emerging as a powerful tool in the establishment of climate science, disaster preparation and interconnected system. Read the highlights in the session or watch the video below.
speaker
Introductory remarks: David SandalowInternational and public schools
» See all Columbia AI Summit video
AI power
“There are many different variations [AI] Possible benefits. What I am particularly excited is in material innovation. 150 years ago, when Thomas Edison invented modern bulbs, he physically tested various types of elements or materials for months, months, and months, and saw how much light and heat would be produced by running electricity through electricity. And he finally came up with what he thought was the best solution. Today, we can simulate millions of these interactions in 2 seconds using the latest AI tools. And that power provides us two advantages. One of them is that you can test materials that do not actually exist. Then, if you look beneficial in a specific application, you can make sure it works and works. And we can choose much faster. The potential for AI to accelerate the speed of energy innovation is large, and you need to find out how to mobilize all kinds of tools. ” –David Sandalow
“What happened now is AI [weather prediction] In essence, models that do not exist as five years ago are good by many indicators, or even some [traditional] Physical -based model. The AI model does not know somewhat about physics in the atmosphere. They do not know energy or momentum preservation or none of them. Like all other AIs, we have been trained in data. You provide them with a lot of historical weather data, know what’s happening in the past, see what’s happening later, and look for patterns in the mysterious way of machine learning and AI. As with other AIs, education is very expensive, but as David said, it is very cheap to run the model in real time in real time, and in a short time, it actually went to a powerful place. ” –Adam Sobel
“I think equality is important. In the past, when there was a problem with the power system (when it occurred during the winter storm URI), the power provider could not supply sufficient energy and reduce demand. During that period, Texas thinks that about 30% of the demand (because I lost a lot of generations because I lost a lot of generations, so I blocked about 30% of the demand, but basically, the easiest way to reduce the demand is to block the branch, but in general, the local community living at the point is generally fewer the latest infrastructure. On the other hand, through the new distribution of batteries and solar PV, we are seeing a rich neighbor who places micro grids such as Tesla Powerwall, which can supply home, and that some families or buildings can supply themselves through micro grids. In this case, there are also areas of advancement to the power system, and sometimes the AI has a lot of promises.Bolun xu
“As a behavioral ecological scholar, I am interested in how individual organisms respond to environmental changes. And in order to start actual predictions, it is necessary to understand the changes of individual levels so that the population and what ecosystems will do according to climate change or land use change. And I think it’s where it is [species identification] Technology can be very useful. We can use images to capture individual zebras by using images in camera traps or pod recording to individual species recognition to individual identification. I think it’s the first species because it has almost barcodes like fingerprints. And you can make a really good population plan, follow the animals, and see the time of drought. To the population and ecosystem, we do better projects on how organisms will react to change. ” –Dustin Ruben Stein
“This is a truly unique opportunity for AI to establish a methodology to determine the optimal adaptation and elasticity strategy for infrastructure for weather and climate -related risks. The problem is extremely challenging, multi -field, and will include engineering, physical sciences, social science and contributions from various fields. And the only way to approach is through the novel. Monte Carlo Including approach Probability. AI will help to integrate these various fields that contribute to the final solution of the problem. ” –George Deodatis
AI challenge
“We know that as long as we predict hurricanes, especially climate, the next 5, 10 or 20 years will cause hurricanes to occur, the frequency and expected intensity are very difficult. And so far, we have some models, probability, which explains this uncertainty. But these models assume that the fixed climate, climate does not change. And since we know that the climate is changing now, this probability of the rectangle for extreme events will change on time. This makes the problem more difficult. So we are relying on artificial intelligence approach to establish some models that can probably probabilistically qualify the development and intensity of these extreme events in the near future. ” –George Deodatis
“If you look at the food system, you can see a lot of inequality. Small farmers, indigenous population and ecosystems, and our food system is often left in this race for better data and decision making. We think we should be very careful about how to use this technology. This is especially because there are many small small farmers in the world. So this data democratization around AI will be very important. ” –Jessica fan
“I have the potential to use things that AI moves so fast and people don’t understand, and now I think there’s a tension between public places and histories. [AI models] Although I came from the private sector, I still rely entirely on the public sector infrastructure on the basic data and the physical models that perform a lot of background. So I think there’s a lot of tension here and I think it’s really dangerous to break things in the infrastructure we depend on to keep people safely. ” –Adam Sobel
“I think about 30%of California’s daily production capacity is now from a huge battery. And it is fair to say that many batteries are operated by AI. Of course with human monitoring… [what the AI is doing]It makes us worry about the power system operator. It is about how to actually understand transparency and regulate this AI. I think this is a real concern for many power system operators. ” –Bolun xu
“We know that the model depends on the data we train, and most people think that we are talking about one servant. Do we want to train a model for one species or one species, then cut the corner and use it to project with other species? I think that’s a real concern. How stylish can the model be to work with other ecosystems or other species regarding the training data? ” –Dustin Ruben Stein
Columbia AI Summit was hosted by Columbia AIA new initiative that promotes Columbia’s work on artificial intelligence with courses, curriculum, events, and digital tools. Columbia AI Data Science Research Institute,,, Columbia Engineeringand Research vice president.
*Highlights have been edited for clarity
