Urban unconcealment

Here is an exchange with ChatGPT, published prematurely it seems, that demonstrates how the LLM seems to have adopted a profile of my interests from uploaded files and previous threads.

Referring to my book “AI and Language in the Urban Context,” I asked: “In the book I make reference to ‘revelation.’ It’s from concepts by Heidegger that as well as ushering in changes, technologies reveal and conceal aspects of our current situation. Is there anything in the AI Index Report 2025 (see previous post) that hints at what LLMs reveal about aspects of our urban lives?”

ChatGPT responded: Yes, there are several points in the AI Index Report 2025 that—though not framed in Heideggerian terms—can be interpreted as moments of disclosure or unconcealment in the sense you explore in the book. While the report doesn’t use the language of revelation or concealment, it inadvertently suggests what LLMs reveal about contemporary urban conditions. Here are three such resonances:

---

1. LLMs Reveal the Centrality of Language in Governance and Public Services
The report notes that governments around the world are increasingly integrating LLMs into public administration, from legal analysis to document drafting. This operational use of language models reveals how thoroughly our bureaucratic and civic systems are constituted by language. As you argue in the book, “cities are negotiated, described, interpreted, analysed and even created through language.” The deployment of LLMs in these contexts makes visible the textual infrastructure of urban governance—something typically backgrounded or taken for granted.

---

2. LLMs Reveal the Extent of Surveillance and Data Dependency
The report flags ongoing concerns about data privacy, bias, and misuse. These concerns arise precisely because LLMs depend on massive, often opaque corpora of data. In your Heideggerian terms, the technology discloses our current reality as one of mass surveillance, commodified language, and algorithmic profiling. It “reveals” an urban condition in which everyday speech and writing are mined, recycled, and re-presented as predictive commodities.

---

3. LLMs Reveal the Prevalence of Narrative Modes of Reasoning
The Index’s recognition that generative models are widely used for content creation, storytelling, and ideation (in fields from education to entertainment) aligns with your claim that cities are structured by narrative practices. These uses of LLMs uncover how much urban rationality already depends on persuasive discourse, not just formal logic or numerical data. This speaks directly to your discussions of Throgmorton and Habermas: AI reveals that persuasive storytelling—not statistical forecasting alone—structures much of our collective urban future-making.

---

In short, the report offers a rich array of indicators that, when read through your interpretive lens, confirm that LLMs don’t just operate in the city—they reveal the city’s linguistic, procedural, and imaginary foundations. These technologies cast familiar practices in a new light, even as they risk obscuring others—an ambiguity very much in line with Heidegger’s view of technē.

As if to diminish further my redundancy in this process, ChatGPT offered to generate a paragraph to fold this “revelation” motif into a blog post draft!

References

• Coyne, R., AI and Language in the Urban Context: Conversational Artificial Intelligence in Cities, London: Routledge, 2025.
• Maslej, N., L. Fattorini, R. Perrault, Y. Gil, V. Parli, N. Kariuki, E. Capstick, A. Reuel, E. Brynjolfsson, J. Etchemendy, K. Ligett, T. Lyons, J. Manyika, J.C. Niebles, Y. Shoham, R. Wald, T. Walsh, A. Hamrah, L. Santarlasci, J. Betts Lotufo, A. Rome, A. Shi and O. Sukrut, The AI Index 2025 Annual Report, Stanford CA: AI Index Steering Committee, Institute for Human-Centered AI, Stanford University, 2025.

Note

• ChatGPT generated the featured image in response to my prompt: Please provide a picture of a broken pencil, concealed in part by a pile of papers with handwriting on them. Evoke a derelict work space.