| 2026 |
“The Humanities in 2056” (American Academy of Arts & Sciences)Categories Talks
|
“The Humanities in 2056,” panel on convening on “What Will We Want to Know about the Humanities in 30 Years?” at convening on “What We Can Know and What We Should Know about the Humanities,” American Academy of Arts & Sciences, 30 October 2026.
Abstract: This presentation began by teasing the future of the humanities through interactive visualizations created by AI (Claude Opus 5.5) to predict how current humanities “nodes”—meaning structures, formations, and concentrations of work in the humanities will develop in relation to other social sectors from the present to thirty years in the future. (See predictive visualizations of the humanities in 2026 and 2056.) In my talk, I proposed that preparing the humanities proactively to plan their own ongoing “restructuring” (instead of just reacting to it) requires thinking about how to migrate existing humanities nodes and create new ones, including allowing for how the overall configuration of the humanities in society will increasingly be steered by media, including AI mediation. Good plannng, though, requires good data about where the humanities are now and where they are heading, which means that humanities data-gathering initiatives such as Humanities Indicators will need to complement their quantitative data-gathering with qualitative data-gathering about what will not easily appear in available quantitative sources: information about emerging academic and extra-academic areas of the humanities. These include, for example, “humanities-plus” programs (such as the medical, legal, environmental, or digital humanities) and—as suggested by Claude—new outside-the-academy humanities work focused on verifying, authenticating, and managing the intellectual property of cultural datra (e.g., for AI training data). Qualitative data-gathering includes ethnographic, documentary, and textual analysis focused on representative cases. To create a bridge between qualitative and quantiative data, materials collected from qualitative study can be interpreted not just through humanities methods but semi-quantitative, structured methods familiar to the social sciences such as grounded theory annotative coding.
- Prompt to several frontier-level AI LLMs at the time of the event (by OpenAI, Google, and Anthropic). Anthropic’s Claude Opus 5.5 was the only AI that produced not a generic result but a deeply interesting interactive visualization representing a prediction that seems to be based on substantive information and data about the humanities in its training datasets. The research question for the humanities: how to produce a version of such a predictive visualization accompanied by analytical and narrative reports that is based on real, knowably-sourced data, both quantitative and qualititative?
- Claude’s visualization of the humanities at present (2026).
- Claude’s visualization of the humanities in 30 years (2056).
Talks