Cracking the code: AI, climate change and policy at the Hertie School

On an early spring afternoon in Berlin, a group of students at the Hertie School were not poring over textbooks or listening passively to a lecture. Instead, they were deep in animated discussions. This lively exchange wasn’t a debate club or a workshop – it was a session of Professor Lynn Kaack’s course, Artificial Intelligence & Climate Change, where solving puzzles isn’t just an exercise in learning, but a key to unlocking new ways of thinking.

The course’s mission

Designed to probe the intricate relationship between two of the most pressing issues of our time – artificial intelligence (AI) and climate change – the course invites students to examine where technology meets policy, and what policymakers can do to align AI tools with environmental goals.

“AI and climate change are two topics that are of great importance – and they are deeply intertwined,” explains Kaack, who was first inspired to dedicate her academic work to climate issues after watching Al Gore’s An Inconvenient Truth as a teenager. “This course isn’t just about energy-hungry data centres. It’s about how AI is applied across the economy – often in ways that may unintentionally hinder addressing our climate goals.”

Puzzling it out, together

The course brings the intersection of AI and climate change to life by having students prepare project presentations, record their own podcast episodes, and prepare input to a mock parliamentary hearing. This week, students worked on a group puzzle assignment.

“The course has made me think seriously about how I can use AI to contribute meaningfully to society.”

For the assignment, students were divided into small groups which each received a different topic. After preparing at home, they first gathered in “homogenous” groups with classmates who had studied the same topic, which allowed them to clarify difficult topics with an in-depth discussion. Later, they rearranged into “heterogeneous” groups to share and discuss their findings with students who had prepared entirely different pieces of the puzzle.

From batteries to biodiversity: Changing perspectives and expanding knowledge

Aditya Narayan Rai, a second-year Master of Data Science for Public Policy (MDS) student, worked on “Machine Learning and Biodiversity”, studying the development of a foundation vision model designed to assist with conservation. “The authors trained a model called BIOCLIP using contrastive learning techniques,” he said. “It allows for species classification even with limited labelled data, which is critical for monitoring biodiversity in real time.”

He says that discussing this in depth with classmates helped solidify his understanding. “Then switching groups gave us exposure to other exciting topics – like reinforcement learning for battery research, and machine learning for electric vehicles,” he added.

Varvara Ilyina, also a second-year MDS student, studied the “Climate Smart Shrimp” case, exploring how AI and satellite data can improve shrimp farming while conserving mangrove forests in Southeast Asia. “Our research showed how computer vision techniques could optimise land use. Explaining this to my fellow classmates pushed me to make complex ideas more accessible.”

Taking lessons out of the classroom

For both students, the exercise did more than teach technical content – it bridged the gap between academic data science and practical, real-world policy application.

“The group changing activity was very helpful in the class,” said Aditya. “Every session links our AI knowledge to climate adaptation or mitigation. I learnt a lot about the diverse topics and how applying machine learning can help address problems related to climate change. And I have more clarity about what skill sets are needed for policy work in the field.”

Varvara echoed the sentiment. “I’ve worked with satellite data before, using AI to find optimal rooftop solar panel placements. This course has shown me how similar tools can be applied across climate-related challenges. It’s made me think seriously about how I can use AI to contribute meaningfully to society after graduation.”

Teaching for tomorrow’s policymakers 

For Kaack, that’s precisely the point. “I want students to come away with the ability to give informed, practical advice to policymakers,” she says. “That means understanding the potential – and the risks – of the technology from both a technical and a systems-level perspective.”

As AI becomes ever more entwined with global sustainability challenges, courses like this one are not just teaching data science or policy – they’re helping train a generation of technologists and thinkers ready to navigate one of the most complex intersections of our time. And if last week’s group puzzle was any indication, these students are already well on their way.

Interested in studying at the Hertie School? Find out more about our programmes: https://www.hertie-school.org/en/study/programmes

Lynn Kaack is Assistant Professor of Computer Science and Public Policy at the Hertie School’s Data Science Lab and Centre for Sustainability. Her research and teaching focuses on methods from statistics and machine learning to inform climate mitigation policy across the energy sector, and on climate-related AI policy.