Stanusch, N., & Rogers, R. (2026). How AI is imagined by industry during the Sam Altman controversy. New Media Soc.. https://doi.org/10.31124/advance.174979411.18178682/v1
Summary
This paper takes the November 2023 firing and rehiring of OpenAI CEO Sam Altman as a critical moment for surfacing how the AI industry imagines and talks about artificial intelligence. Drawing on controversy mapping and cross-platform digital methods, Stanusch and Rogers identify three dominant industry-aligned imaginaries—Longtermism, Regulatory Ambivalence, and Techno-Hagiography—and argue that the industry uses two rhetorical strategies to manage the concerns these imaginaries raise: premediation (externalizing critique into a speculative future) and preclusion (internalizing critique by casting itself as the actor best able to resolve it). Together these tactics let the industry “occupy the future and absorb the present,” though counter-imaginaries circulating on X/Twitter show this monopolization is incomplete.
Key Contributions
- Introduces a framework treating AI imaginaries as sites of issuefication—actively shaping which concerns become salient and which are rendered invisible.
- Identifies premediation and preclusion as concrete tactics by which the AI industry manages controversy.
- Provides an empirical, cross-platform mapping of AI imaginaries during a discrete, high-visibility controversy.
- Extends controversy mapping and digital methods to AI industry self-representation on professional and discourse-oriented platforms.
- Demonstrates platform specificity: LinkedIn as an aspirational space reinforcing industry imaginaries versus X/Twitter as a more contested space hosting counter-imaginaries.
Methods
The authors use Latourian controversy mapping, treating the Altman incident as an opportune moment to trace actor associations. They scraped ~1,900 LinkedIn posts (via queries on Sam Altman) and ~1,700 X/Twitter posts (from 101 AI-related organizations and associated accounts) covering Nov 2023–Jan 2024, using the Zeeschuimer extension and research accounts due to API closures. Network analysis in Gephi (co-hashtag and mention networks for LinkedIn, retweet networks for X/Twitter) was combined with the 4CAT toolkit and a three-step inductive qualitative coding process—ethnographic content analysis, keyword/actor/issue identification, and labeling of emergent imaginaries. Post content was anonymized through paraphrasing.
Findings
- On LinkedIn, three imaginaries dominated: Longtermism (AGI, superalignment, existential risk), Techno-Hagiography (mythologizing Altman as a lone-genius hero), and Regulatory Ambivalence (acknowledging regulation while resisting external oversight).
- LinkedIn also featured “LinkedIn Talk” (self-promotion, recruitment) and cryptocurrency discourse leveraging the controversy.
- On X/Twitter, counter-imaginaries emerged: Governing Urgency (demanding strong governmental regulation) and Intersectionality (highlighting white-male dominance, racism, and anti-Palestinian bias, including the Tal Broda controversy).
- An imaginary of the Technological Sublime emerged around copyright and the black-boxed unpredictability of AI capability.
- Present-day concerns—environmental costs, labor implications, slow-AI alternatives, historical comparisons—were notably absent from the discourse.
- Longtermism’s roots trace to early-2000s effective altruism and AI Safety communities that displace present harms with speculative future threats.
Connections
This paper sits within the platform-critique lineage that treats industry self-representation and imaginaries as objects of critical scrutiny; its concern with how AI discourse is framed and contested connects it to work on AI’s epistemic and public effects such as Lewandowsky2026-ob and Bak-Coleman2026-mk. Its methodological reliance on cross-platform digital methods and scraping under API closures resonates with infrastructural and data-access critiques like Bruns2026-yv and Helmond2026-ll.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts