High-powered Reference Access to Text

Title: High-powered Reference Access to Text

Description: 

In the world of AI, archival repositories face increasing interest in leveraging technology to enhance research access to text. For example, researchers might want to apply keyword searching, sentiment analysis, or other tools to groups of document files. Archivists might want to develop deeper descriptive access in finding aid notes such as Scope and Contents. Providing meaningful research access to large collections of born-digital or digitized texts often remains challenging due to limited staffing, resources, and time.

This half-day workshop will explore ways in which archivists can approach enhancing reference for text using options suitable for a range of institutional contexts and technical comfort levels. It will provide practical, scalable methods for leveraging computational tools through both non-AI and AI-assisted approaches. This includes basic command-line scripts, spreadsheet-based analysis, and the option of AI tools for further enhancement. Tasks which can be accomplished using command-line formulas are freely accessible to anyone and do not require special software or AI. Results from these command-line formulas can be examined in Excel and further developed, for example with charts and graphs. In addition, more nuanced data analysis can be developed with AI tools if desired. Through discussion and hands-on exercises using participants’ own laptops, attendees will gain frameworks to evaluate when and how to incorporate computational methods for access to archival texts. We will address ethics and vulnerabilities of using AI, such as environmental impact, data security, and the potential for data hallucinations. Participants will leave with adaptable workflows, sample templates, and the ability to incorporate skills learned in their own work.

Participation in the hands-on portion is optional. The learning objectives can also be accomplished by observing the demonstrations we provide.

Summary:
Format: Small, classroom-style (15 participants max)
Target Audience: Cultural workers, archivists, historians, genealogists, and those working with archival materials
Learning objectives: 

1. Identify common challenges to research access in archival text collections and assess when computational approaches might be useful
2. Create .txt files from letters and other documents
3. Apply basic command-line and spreadsheet techniques to search, group, and analyze texts
4. Explore AI-assisted text analysis methods and compare with non-AI alternatives, understanding the strengths and limitations of each 
5. Evaluate ethical and professional considerations if/when using AI based tools to enhance archival workflows 
6. Design a potential workflow suitable for your institutional context
Hands-on components: Participants are welcome to attend with their own laptop and try some of the techniques we will use hands-on, or they may attend with only the intention to listen and ask questions, but not engage hands-on. No prior command-line experience is required.

Schumacher Flex Commons, Moody Memorial Library, Baylor University

1312 South 3rd Street

Waco, TX 76706

(Further directions will be emailed to registrants)

Time: 8:30 AM to 12:30 PM

Capacity: 15 in-person participants

Required Knowledge and/or Equipment: No prior command-line experience is required.

Proposed Workshop AV and Materials Required: Projection screen for PowerPoint slides.

Instructors:  Megan Scott and Amanda Focke 
Email: ms531@rice.edu; afocke@rice.edu

Refunds: Full refunds for annual meeting and workshop registrations are available no less than 10 business days prior to the first day of the annual meeting. Refunds requested after that date will incur a $25 penalty.