GSAPP Faculty Retreat
Adam Vosburgh
GSAPP AI faculty retreat
September 18, 2026
Slides, sources, further reading
AI traces in GSAPP work
Slides
Link to slides (PDF, 41 pages, 4.5 MB)
The screen recordings in the deck appear in the PDF as still frames.
Mentioned in the presentation
Methods in Spatial Research, "Making Spatial Data from Plain Text" Tutorial , Columbia GSAPP, Spring 2026
Kate Crawford, Atlas of AI , 2021
Touvron et al., "LLaMA: Open and Efficient Foundation Language Models" , Meta AI, 2023
Common Crawl
C4 ; Raffel et al., 2019
CompVis, "Stable Diffusion v1-4" model card , 2022
Schuhmann et al., "LAION-5B: An open large-scale dataset for training next generation image-text models" , NeurIPS 2022
LAION, "Releasing Re-LAION 5B" , 2024
Christo Buschek and Jer Thorp, Models All The Way , Knowing Machines, 2024
Kousha and Thelwall, "How much are LLMs changing the language of academic papers after ChatGPT?" , Scientometrics , 2026; preprint arXiv:2509.09596
Kobak et al., "Delving into LLM-assisted writing in biomedical publications through excess vocabulary" , Science Advances , 2025; preprint arXiv:2406.07016
Chum, Krisch, Kurgan, Vosburgh, Zhang, The Curse of Dimensionality , Venice Architecture Biennale, 2025 Project we worked on about the use of AI models in conflict. The last chapter attempts to visualize the high-dimensional space of a targeting machine.
Stable Diffusion 3.5 Medium , Stability AI, 2024
Claude Code , Anthropic
MS CDP, Statement on Coding & AI: Situated Computation in the Age of Assistants , Columbia GSAPP, 2025
Simulations, Models, Twins , Columbia GSAPP, Fall 2026
Further reading
Karen Hao, Empire of AI: Dreams and Nightmares in Sam Altman's OpenAI , Penguin, 2025
Rebecca Tan and Regine Cabato, "Behind the AI boom, an army of overseas workers in 'digital sweatshops'" , The Washington Post , August 2023
Kyle Barr, "Anti-Piracy Group Takes AI Training Dataset 'Books3' Offline" , Gizmodo , August 2023
Hans et al., "Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text" , 2024 Paper describing a method of AI detection using perplexity (how surprising text is when read by a LLM) adjusted by how formulaic the text is expected to be (e.g. heavily templatized text is always going to be less surprising.)
Oviedo et al., "Energy Use of AI Inference, Efficiency Pathways, and Test-Time Scaling" , Joule , 2026; preprint arXiv:2509.20241
"Silicon Valley Splits Over Closing the Borders to Chinese A.I." , The New York Times , July 25, 2026
Yuval Abraham, "'Lavender': The AI machine directing Israel's bombing spree in Gaza" , +972 Magazine and Local Call, April 3, 2024
Anthropic, "Measuring AI capabilities in intelligence targeting and conventional weapons" , September 10, 2026In this report, there are a couple of very illuminating examples of how AI models can turn Signals Intelligence (SIGINT) into precise locations (often for airstrikes.)
Oliver Whang, "We Don't Really Know How A.I. Works. That's a Problem." , The New York Times Magazine , April 15, 2026
AlphaEarth Foundations , Google DeepMind, July 2025 Embeddings models are a new type of AI model that (at a very high level) use satellite data instead of text. Truly bizarre in many ways, and difficult to understand the implications of.
Ellie Madsen, Embedding Earth: A Critical Atlas of Planetary Computation , 2026 2026 CDP Capstone project attempting to make sense of embeddings models.
We Asked an AI to Map Our Stories Across NYC , The City, February 2024
Genie 3 , Google DeepMind, August 2025
CDP Capstone projects
GSAPP EOYS projects