Generative Artificial Intelligence and the future of Digital Humanities in Europe
Generative Artificial Intelligence and the future of Digital Humanities in Europe
Davit Sidamonidze
Independent Researcher / University of Warsaw
ORCID: 0000-0002-0386-896X
Nana Deisadze
Independent Researcher / Tbilisi State University
ORCID: 0000-0003-0561-1719
Abstract
The emergence of generative artificial intelligence (GenAI) is reshaping digital humanities research, offering new possibilities for textual analysis, cultural heritage studies, multilingual scholarship, and computational interpretation. Large language models and generative systems increasingly support researchers in corpus analysis, translation, metadata generation, and scholarly communication. However, these developments also raise concerns regarding interpretation, authorship, transparency, and research integrity. This paper examines the implications of generative AI for European digital humanities. Drawing upon recent developments in computational humanities, AI governance, and European research infrastructures, the paper analyzes opportunities and challenges associated with AI integration. Particular attention is given to multilingual corpora, cultural heritage, historical archives, and scholarly interpretation. The study proposes principles for responsible AI use within digital humanities research and argues that AI should function as an interpretive assistant rather than a replacement for human scholarship.
Keywords: digital humanities, generative AI, cultural heritage, computational humanities, Europe, research integrity
- Introduction
Digital humanities have transformed humanities scholarship through computational methods, large-scale textual analysis, and digital archives. During the past two decades, techniques such as topic modeling, text mining, network analysis, and corpus linguistics have expanded the methodological toolkit of humanities researchers.
Generative artificial intelligence represents the newest stage of this transformation.
Large language models can:
- summarize texts,
- generate metadata,
- support translations,
- identify themes,
- assist coding,
- facilitate multilingual analysis.
European digital humanities increasingly operate within large infrastructures such as:
- CLARIN,
- DARIAH,
- Europeana,
- national digital archives.
The integration of AI within these infrastructures presents both opportunities and risks.
This paper asks:
- How can generative AI support digital humanities?
- What challenges emerge for interpretation and authorship?
- How can responsible AI support humanities scholarship?
- Digital Humanities and Computational Interpretation
Digital humanities emerged from the application of computational methods to humanities questions (Berry, 2012).
Methods include:
- corpus analysis,
- topic modeling,
- network analysis,
- geospatial humanities,
- digital archives.
Jockers (2013) introduced macroanalysis as a means of examining literary and cultural patterns at large scales.
However, humanities scholarship remains fundamentally interpretive.
Underwood (2019) argues that computational methods should support rather than replace human interpretation.
Generative AI intensifies this debate.
- AI Applications in Digital Humanities
3.1 Textual analysis
AI can support:
- topic extraction,
- summarization,
- semantic analysis,
- entity recognition.
Large corpora become more accessible.
3.2 Multilingual research
Europe’s linguistic diversity creates challenges.
Generative AI enables:
- translation,
- cross-linguistic comparison,
- multilingual retrieval.
This may strengthen comparative humanities.
3.3 Cultural heritage
Museums and archives increasingly use AI for:
- metadata generation,
- digitization,
- cataloguing.
Digital heritage institutions benefit from automation.
3.4 Historical archives
Historical documents often contain:
- OCR errors,
- incomplete metadata,
- linguistic variation.
AI may improve accessibility.
- Challenges
Several concerns emerge.
4.1 Authorship
Who authors AI-assisted interpretations?
Human scholars remain responsible.
4.2 Hallucinations
Generative AI may produce inaccurate information.
Verification is essential.
4.3 Loss of interpretation
Humanities scholarship depends on:
- context,
- meaning,
- critique.
AI cannot replace these functions.
4.4 Bias
Training datasets may reproduce cultural biases.
European multilingual diversity remains unevenly represented.
- European Research Infrastructures
European digital humanities benefit from:
- CLARIN,
- DARIAH,
- Europeana,
- EOSC.
These infrastructures provide opportunities for:
- shared guidelines,
- repositories,
- training.
AI governance should become integrated into digital humanities infrastructures.
- Responsible AI for Digital Humanities
Four principles are proposed:
- Transparency.
- Human interpretation.
- Documentation.
- Ethical responsibility.
AI should augment scholarship.
It should not replace scholarly judgment.
- Discussion
Generative AI may democratize digital humanities.
Smaller institutions can access advanced tools.
Researchers from peripheral regions may participate more effectively.
However, unequal access and varying digital capacities may create new disparities.
European cooperation can reduce these risks.
- Conclusion
Generative AI represents a major transformation for digital humanities.
Its greatest contribution lies not in replacing interpretation but in supporting new forms of inquiry.
European digital humanities can play a leading role in developing responsible, transparent, and inclusive AI practices.
References
Berry, D. M. (2012). Understanding digital humanities. Palgrave Macmillan.
Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet allocation. Journal of Machine Learning Research, 3, 993–1022.
Jockers, M. (2013). Macroanalysis. University of Illinois Press.
Underwood, T. (2019). Distant horizons. University of Chicago Press.
European Commission. (2020). A European strategy for data.
UNESCO. (2021). Recommendation on the ethics of artificial intelligence.