AI AND THE DIGITAL HUMANITIES

In Digital Humanities (DH), AI techniques are used to analyze large amounts of digitized cultural data, like texts, images, and audio, to identify patterns, extract meaning, and gain new insights into human history and culture, allowing researchers to explore complex topics at a scale previously impossible through manual analysis.

Specific examples of AI techniques in DH include:

Audio analysis:

• Speech recognition: Transcribing audio recordings of historical speeches or interviews. 
• Music analysis: Identifying musical patterns, styles, and composers within audio recordings. 

Data visualization:

• Interactive maps: Creating dynamic visualizations of geographical data related to historical events or cultural phenomena. 
• Network graphs: Representing relationships between people, places, or concepts within a dataset.

Digital preservation:

• Metadata extraction: Automatically generating descriptive metadata for digital artifacts.
• Data cleaning and normalization: Standardizing inconsistent data to improve analysis capabilities.

Historical data analysis: 

• Extracting information from large datasets of historical documents, mapping geographic trends.
• Mapping the movement of people or ideas across regions: using geographical data and text analysis. 

Image analysis:

• Object recognition: Automatically identifying objects within historical images (e.g., buildings, artifacts. historical photographs, and paintings). 
• Facial recognition to identify individuals in historical records. 
• Image captioning: Generating descriptive text for images to improve accessibility.
• Image restoration: Image restoration and enhancement of degraded digital artifacts (Digitally repairing damaged historical photographs or artwork) .

Sentiment analysis: Assessing the emotional tone of historical documents. 

Text analysis: 

• Identifying recurring themes and motifs in large literary corpora, authorship attribution to determine the writer of unknown texts.
• Analyzing the evolution of language in historical texts by identifying changes in word usage and grammatical patterns over time. 

Stylometry: stylistic analysis of literary works; identifying the likely author of a text by analyzing writing style patterns. 

Topic modeling: Discovering recurring themes and topics within large collections of texts.

Some early examples of AI applications in humanities research, such as projects analyzing ancient texts using computational methods, can be traced back to the 1960s and 1970s. With the rise of deep learning and large language models, AI has become much more integrated into various Digital Humanities practices, including text mining, sentiment analysis, and automated translation. 

A.  Digital Palaeography, is a notable example of early AI application in humanities. It is the study of historical handwriting using digital tools (computer algorithms) to analyze and interpret ancient or antiquated writing systems, allowing for more comprehensive comparisons and analysis of manuscripts through digital databases and computational methods, compared to traditional manual palaeographic techniques. 

While the exact date is difficult to pinpoint, the emergence of Digital Palaeography as a distinct field is generally considered to have begun in the late 20th century with the increased availability of digital imaging technology for manuscripts, alongside the development of specialized software and databases to analyze them, with significant advancements occurring particularly in the early 21st century; notable examples include the "D-scribes" project which started in 2018 focusing on digital analysis of Greek and Coptic papyri.  (For Project Summary click here).

B.  Considered one of the first AI systems to produce creative outputs in the art world, highlighting the early exploration of AI in artistic expression is the "AARON" program, developed by artist Harold Cohen. AARON was a single program that involved teaching a robot to create drawings. Aaron’s education took a similar path to that of humans, evolving from simple pictographic shapes and symbols to more figurative imagery, and finally into full-color images. Aaron was capable of generating original artwork, demonstrating the potential for AI to creatively mimic artistic styles and techniques within the realm of visual arts. It could generate drawings and paintings by simulating artistic decisions like line quality, color selection, and composition based on a set of rules and parameters. AARON's development paved the way for further research into AI-generated art, demonstrating the potential of computational tools to mimic and even expand artistic possibilities.