The integration of AI with Climate Science offers transformative capabilities for understanding and combating climate change. AI’s capacity to analyze vast datasets—from historical weather patterns to real-time satellite imagery—can significantly advance climate modeling, impact assessment, and the development of mitigation strategies. However, the deployment of AI in this field must be navigated with a clear understanding of both the ethical implications and the need for scientific rigor.
Examine the ethical and scientific implications of utilizing a scientific AI training dataset or AI process used in climate science with applications toward scientific discovery.
If you choose to focus on training data, you may choose multiple examples of AI training sets to support your reasoning or focus on a single example. The data must be publicly available and published by a reputable source. Judges should be able to access the data set(s).
If you choose to focus on AI algorithms or processes, detail the steps taken for your analysis of how the process was engineered and how it is used.
Choose a particular ethical framework for your discussion.
Optional: Demonstrate an AI model trained with the dataset to support arguments for responsible and accurate AI applications in scientific research.
Your submission should be crafted for a general audience as much as possible. Remember that we are open to creative formats for your work.
Guidance
We need to think early and often about the potential biases inherent in AI algorithms and training data. Ultimately you need to explore what needs to be done to demonstrate that the AI processor training data can be trusted to deliver rigorous and ethical science over time. You need to define specifically what you mean by rigorous and ethical.
The suggestions below are not meant to be prescriptive but are there to help you think about how to frame your discussion.
Background Material
Tackling Climate Change with Machine Learning: https://mitsloan.mit.edu/ideas-made-to-matter/tackling-climate-change-machine-learning