Generative AI tools, such as ChatGPT, simulate human writing and complicate the inquiry-driven culture we live in. These tools use singular first-person pronouns in their textual outputs and are often associated with anthropomorphic qualities. Within the humanities, conversations tend to focus on detecting new forms of plagiarism. What is missing from the current debate are insights from performance studies. As a synthesis of human-generated datasets, generative AI is changing publics' relationship to themselves.

Since ChatGPT remixes statistically most likely combinations of words, its outputs are in fact a form of theatrical performance. It draws on users' prompts and the publics' collective memories to produce improvised performances, within specific parameters, for its user-audiences. As a mirror held up to the humanity, ChatGPT produces a pixelated shadow of the publics in time.

If we take ChatGPT to 500 BC, it would insist, as the society did, that the Earth is flat. ChatGPT is therefore a survey instrument of the publics' collective biases rather than the truths. It is an aesthetic instrument rather than an epistemological tool. Based on this understanding, this interactive presentation will theorize AI in the framework of digital humanities and provide pedagogical strategies for educators to teach with AI rather than against it.