Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.16403 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911602876350464 |
|---|---|
| author | Kommers, Cody Ahnert, Ruth Antoniak, Maria Benetos, Emmanouil Benford, Steve Bunz, Mercedes Caramiaux, Baptiste Concannon, Shauna Disley, Martin Dobson, James Du, Yali Duéñez-Guzmán, Edgar Francksen, Kerry Gius, Evelyn Gray, Jonathan W. Y. Heuser, Ryan Immel, Sarah So, Richard Jean Leigh, Sang Livingston, Dalaki Long, Hoyt Martin, Meredith Meyer, Georgia Mihai, Daniela Noel-Hirst, Ashley Ostherr, Kirsten Parker, Deven Qin, Yipeng Ratcliff, Jessica Robinson, Emily Rodriguez, Karina Sobey, Adam Underwood, Ted Vashistha, Aditya Wilkens, Matthew Wu, Youyou Zheng, Yuan Hemment, Drew |
| author_facet | Kommers, Cody Ahnert, Ruth Antoniak, Maria Benetos, Emmanouil Benford, Steve Bunz, Mercedes Caramiaux, Baptiste Concannon, Shauna Disley, Martin Dobson, James Du, Yali Duéñez-Guzmán, Edgar Francksen, Kerry Gius, Evelyn Gray, Jonathan W. Y. Heuser, Ryan Immel, Sarah So, Richard Jean Leigh, Sang Livingston, Dalaki Long, Hoyt Martin, Meredith Meyer, Georgia Mihai, Daniela Noel-Hirst, Ashley Ostherr, Kirsten Parker, Deven Qin, Yipeng Ratcliff, Jessica Robinson, Emily Rodriguez, Karina Sobey, Adam Underwood, Ted Vashistha, Aditya Wilkens, Matthew Wu, Youyou Zheng, Yuan Hemment, Drew |
| contents | Generative AI systems are increasingly recognized as cultural technologies, yet current evaluation frameworks often treat culture as a variable to be measured rather than fundamental to the system's operation. Drawing on hermeneutic theory from the humanities, we argue that GenAI systems function as "context machines" that must inherently address three interpretive challenges: situatedness (meaning only emerges in context), plurality (multiple valid interpretations coexist), and ambiguity (interpretations naturally conflict). We present computational hermeneutics as an emerging framework offering an interpretive account of what GenAI systems do, and how they might do it better. We offer three principles for hermeneutic evaluation -- that benchmarks should be iterative, not one-off; include people, not just machines; and measure cultural context, not just model output. This perspective offers a nascent paradigm for designing and evaluating contemporary AI systems: shifting from standardized questions about accuracy to contextual ones about meaning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_16403 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Computational Hermeneutics: Evaluating generative AI as a cultural technology Kommers, Cody Ahnert, Ruth Antoniak, Maria Benetos, Emmanouil Benford, Steve Bunz, Mercedes Caramiaux, Baptiste Concannon, Shauna Disley, Martin Dobson, James Du, Yali Duéñez-Guzmán, Edgar Francksen, Kerry Gius, Evelyn Gray, Jonathan W. Y. Heuser, Ryan Immel, Sarah So, Richard Jean Leigh, Sang Livingston, Dalaki Long, Hoyt Martin, Meredith Meyer, Georgia Mihai, Daniela Noel-Hirst, Ashley Ostherr, Kirsten Parker, Deven Qin, Yipeng Ratcliff, Jessica Robinson, Emily Rodriguez, Karina Sobey, Adam Underwood, Ted Vashistha, Aditya Wilkens, Matthew Wu, Youyou Zheng, Yuan Hemment, Drew Artificial Intelligence Computers and Society Generative AI systems are increasingly recognized as cultural technologies, yet current evaluation frameworks often treat culture as a variable to be measured rather than fundamental to the system's operation. Drawing on hermeneutic theory from the humanities, we argue that GenAI systems function as "context machines" that must inherently address three interpretive challenges: situatedness (meaning only emerges in context), plurality (multiple valid interpretations coexist), and ambiguity (interpretations naturally conflict). We present computational hermeneutics as an emerging framework offering an interpretive account of what GenAI systems do, and how they might do it better. We offer three principles for hermeneutic evaluation -- that benchmarks should be iterative, not one-off; include people, not just machines; and measure cultural context, not just model output. This perspective offers a nascent paradigm for designing and evaluating contemporary AI systems: shifting from standardized questions about accuracy to contextual ones about meaning. |
| title | Computational Hermeneutics: Evaluating generative AI as a cultural technology |
| topic | Artificial Intelligence Computers and Society |
| url | https://arxiv.org/abs/2604.16403 |