The AI Model Risk Catalog: What Developers and Researchers Miss About Real-World AI Harms
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911117363642368 |
|---|---|
| author | Rao, Pooja S. B. Šćepanović, Sanja Jayagopi, Dinesh Babu Cherubini, Mauro Quercia, Daniele |
| author_facet | Rao, Pooja S. B. Šćepanović, Sanja Jayagopi, Dinesh Babu Cherubini, Mauro Quercia, Daniele |
| contents | We analyzed nearly 460,000 AI model cards from Hugging Face to examine how developers report risks. From these, we extracted around 3,000 unique risk mentions and built the \emph{AI Model Risk Catalog}. We compared these with risks identified by researchers in the MIT Risk Repository and with real-world incidents from the AI Incident Database. Developers focused on technical issues like bias and safety, while researchers emphasized broader social impacts. Both groups paid little attention to fraud and manipulation, which are common harms arising from how people interact with AI. Our findings show the need for clearer, structured risk reporting that helps developers think about human-interaction and systemic risks early in the design process. The catalog and paper appendix are available at: https://social-dynamics.net/ai-risks/catalog. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_16672 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The AI Model Risk Catalog: What Developers and Researchers Miss About Real-World AI Harms Rao, Pooja S. B. Šćepanović, Sanja Jayagopi, Dinesh Babu Cherubini, Mauro Quercia, Daniele Computers and Society Artificial Intelligence We analyzed nearly 460,000 AI model cards from Hugging Face to examine how developers report risks. From these, we extracted around 3,000 unique risk mentions and built the \emph{AI Model Risk Catalog}. We compared these with risks identified by researchers in the MIT Risk Repository and with real-world incidents from the AI Incident Database. Developers focused on technical issues like bias and safety, while researchers emphasized broader social impacts. Both groups paid little attention to fraud and manipulation, which are common harms arising from how people interact with AI. Our findings show the need for clearer, structured risk reporting that helps developers think about human-interaction and systemic risks early in the design process. The catalog and paper appendix are available at: https://social-dynamics.net/ai-risks/catalog. |
| title | The AI Model Risk Catalog: What Developers and Researchers Miss About Real-World AI Harms |
| topic | Computers and Society Artificial Intelligence |
| url | https://arxiv.org/abs/2508.16672 |