SESSION: AI in the Art History Classroom

Artificial Intelligence is reshaping how we research, analyse, and teach art history. As generative models, image synthesis tools, and data-driven analytics become increasingly embedded in educational practice, their use in disciplines traditionally grounded in human creativity, critical interpretation, and visual literacy raises both exciting opportunities and serious concerns. This panel explores the pedagogical and ethical implications of integrating AI technologies into art history instruction.

We invite papers that critically engage with the impact of AI tools—such as DALL·E 2, Midjourney, Stable Diffusion—on teaching and learning. Can AI foster creativity, enhance critical and analytical skills, or support student engagement? How do we responsibly address the risks of bias, plagiarism, misinformation (“hallucinations”), and the unauthorized use of authors’ and artists’ work in training datasets? How should we assess student work when AI-generated content may be indistinguishable from human-authored assignments?

This session aims to facilitate dialogue around best practices for integrating AI into the classroom. We welcome submissions exploring any aspect of AI’s impact on art history education, including but not limited to:

Using virtual and augmented reality to deliver experiential art history instruction.
Equipping students with the skills to critically evaluate and interpret AI-generated content.
Partnering with museums to develop AI-powered digital exhibitions and educational tools.
Designing coursework that incorporates AI tools to support creativity and critical thinking.
Developing assessment strategies that prepare students for careers in AI-influenced professional contexts.
Ensuring that AI-enhanced learning remains inclusive, accessible, and sensitive to diverse student needs.

Session Convenors:

Lenia Kouneni, University of St Andrews

Natalia Sassu Suarez Ferri, University of St Andrews

Speakers:

Helena Schmidt, Institute for Art Education, Academy of Fine Arts Vienna

Hallucination, Imagination and Digital Dreaming. Bias, In/Justice and Toxicities in Digital Art Education

This paper presentation examines anthropomorphic powers attributed to AI tools, such as dreaming, imagining and hallucinating. This attribution is reinforced by early AI interface commands like “Imagine” and tools like “Deep Dream,” which frame complex computational processes as mystical, unpredictable events. However, digital dreams can quickly become nightmares when AI hallucinates ‘facts’ or is weaponized to promote harmful agendas. Teaching digital visual literacy is essential to empower students concerning AI-generated texts and visuals. LLMs and GenAI image software promote bias, injustices and toxicities deeply inscribed in their training data and reflect the power of those who finance and build them.

This paper proposes critical strategies for art education that positions AI as a site of critical inquiry, where algorithmic bias, data politics and creative agency collide. How might AI’s ‘hallucinations’ (errors, glitches, biases, dreams) reveal systemic inequities and opportunities for subversion? How can we critique the neoliberal framing of AI as a creative efficiency tool? By learning to prompt ‘otherwise’ (Krasny, Krejs 2024; Olufemi 2021), and ‘unlearn’ algorithmic bias (Spivak 1996, Sternfeld 2014), we can reimagine critical AI literacies through anti-oppressive pedagogies. My presentation offers actionable classroom strategies drawing on artistic case studies that critically engage with AI and digital dreaming to subvert colonial, heteropatriarchal, ableist, and extractivist paradigms.

Tiange Zhou, Beijing Normal University School of Future Design

Algorithmic Empathy: Affective Computing as a Pedagogical Tool for Cross-Cultural Analysis in Art History

This paper addresses the pressing challenges posed by generative artificial intelligence to art history pedagogy by proposing ‘algorithmic empathy’ as a critical textual and cultural methodology. Diverging from prevalent applications in image synthesis, this research redirects affective computing towards the domain of art texts themselves—including criticism, exhibition catalogues, and curatorial essays. By analysing how emotion is linguistically encoded within English- and Chinese-language art-historical discourses, the study contends that this approach functions as a tool for comparative cultural analysis, objectively revealing the culturally specific value systems that underpin aesthetic judgement.

Grounded in Geert Hofstede’s cross-cultural value theory and the framework of computational hermeneutics, the investigation constructs a corpus of modern and contemporary art texts. Transformer-based sentiment models (e.g., multilingual BERT) are employed to extract patterns of evaluative affect embedded within critical writing. The results are interpreted not as psychological data, but as ‘cultural signatures of feeling’—distinctive ways in which linguistic traditions encode and express empathy, hierarchy, or transcendence through aesthetic language.

Pedagogically, ‘algorithmic empathy’ is advanced as a means of cultivating critical affective literacy. By guiding students to compare affective vocabularies and their algorithmic interpretations across cultural contexts, this framework equips them to analyse how AI functions as a refractor of cultural emotion. This process enables students to interrogate AI’s role in representing and structuring cultural perspectives. All in all, this paper argues that textual affective computing can enrich art history education by transforming emotion into a site of intercultural analysis—quantified, visualised, and critically reinterpreted—thereby providing a methodological innovation for the discipline.

Samantha Chang, University of Toronto

Making Connections with GenAI: Critical Creativity in the Art History Classroom

This paper presents a pedagogical experiment integrating generative AI into art history teaching as a catalyst for critical creativity and synthesis. In an undergraduate course on seventeenth-century art and architecture, students complete the ‘Making Connections Essay’, responding to GenAI-generated images created from course materials—artworks, readings, and lecture content. The task is not to analyse the AI itself, but to use these hybrid images as prompts for connecting ideas, objects, and themes across the semester, bringing together formal, comparative, and contextual methods of analysis.

Developed through iterative, reflective teaching practice, the assignment positions GenAI’s distortions as productive rather than problematic. Its open-ended structure encourages curiosity, sustained engagement, and authentic learning closely tied to course content. The approach reduces concerns about academic integrity while fostering deeper, more synthetic writing and discussion. For instructors, the process becomes an occasion for reflexive critique, revealing both the algorithmic and disciplinary biases that shape our understandings of art and history.

By treating GenAI as a partner in enquiry rather than a substitute for insight, this work demonstrates how art history pedagogy can engage emerging technologies ethically and creatively. The ‘Making Connections Essay’ offers a scalable model for cultivating critical, creative, and inclusive learning in an increasingly algorithmic visual world.

Barbara Tramelli, Free University of Bozen

From Prompt to Picture: Exploring AI-Driven Creativity and Its Implications in the Classroom

The pervasiveness of digital technologies is driving a profound transformation in the domains of drawing and visual representation, marked by the shift from analog to digital media, the integration of AI-based methods, and the emergence of new visual languages. In this context, studying how creativity interfaces with AI systems endowed with considerable autonomy in image production is essential to understanding the role of the author in guiding and governing the tool’s poiesis, both inside and outside the classroom. This paper offers a theoretical framing of text-to-image applications within the broader landscape of art history and visual production, providing a clear explanation of their operational mechanisms. Drawing on experiences from courses in Digital Humanities at the University of Bolzano, it explores how students not only engage in the production of images and content but also actively reflect on the evolving AI-based media, considering the epistemic implications of AI-generated representation. Text-to-image tools, such as DALL·E and Stable Diffusion, are positioned as central to the creation of new representational practices, with student feedback serving as a valuable resource for understanding the creative process and the dynamics of human-AI collaboration in visual storytelling and educational settings.