This summer, I digitized over 180 cases from the 1980 Fairfax County Circuit Court chancery, civil, and criminal dockets for Dr. Robertson’s course in the spring. Though these digitized cases don’t exist in a searchable, publicly accessible database, this week’s readings gave me an idea of what it would be like to turn them into one. Here are three things I learned:
- I am impressed by the sheer amount of work that goes into digitizing large collections for the public. The process of gathering the “big data” for my project- unbinding, scanning, cropping, and rebinding tens of thousands of pages took a significant amount of time, labor, and money. Turning the raw documents into “smart data,” or transcribing and creating searchable metadata for them all, would require exponentially more of these resources. I will discuss AI transcription later in this post, but I agree that Schoch’s proposed solutions of automation and crowdsourcing would be a more affordable alternative.1
- With my limited resources, I gained firsthand experience as an active historical editor, parsing and gutting large piles of preserved documents to find certain historical data. As the digitizer, I strongly agree with Ziegler’s argument that “records are not neutral,” as I am aware of conscious decisions I made that shaped how the cases exist in a digital format.2 I was instructed to exclude certain records and their information, including sheriff summons, documents regarding court fees and their status, etc. The editing made sense for the task at hand, but left out a wealth of valuable data—for researchers looking into 1980s court debt, for instance. If the court cases were to be made publicly accessible, they would surely need to be re-scanned with a wider scope of inclusion.
- During the process, I found myself wrestling with the ethical implications of digitization, particularly when digitizing sensitive cases. Despite access being limited to a small group of researchers, the archive required certain redactions on digital copies to protect individuals’ privacy. All Social Security numbers, along with minors’ names, birth dates, and faces in the digitized case files, are now permanently covered with a black rectangle—a necessity made even more important due to the recency of the cases. I was also restricted from digitizing medical records or anything marked as confidential. These were the only restrictions; I was free to digitize graphic descriptions and images of violence and assault, personal testimony, addresses, and financial information. After reading Ziegler’s article on the privacy issues that accompany open data projects, I ultimately think that if the county’s court cases were digitized for open use, access to them should be limited in some way out of respect for the human subjects involved. This limitation could include a hard cutoff of public access to all digitized court records at a certain year (1926, for example), more comprehensive redactions on publicly accessible copies of these documents, individual approval requirements for researchers, or something else entirely.
Digital Research and Databases
As a frequent user of the LOC’s Chronicling America and D.C. Public Library’s zine collections, I find digitized historic databases to be incredibly useful tools for research. Library and archival digitization projects increase access to historical information, removing barriers to research including cost, distance, and, as Anna pointed out during our group discussion, health concerns. However, numerous authors this week identified problems with using digitized sources instead of in-person materials for historical research. Lara Putnam offers two: first, that keyword search functions can give historians “cherry-picked” evidence instead of the historical analysis that would traditionally come from engaging with sources in an “immersive” way, and that researchers can lose out on crucial cultural and local knowledge by accessing the materials out of their original archival contexts. These are problems that arise from the problematic use of these digital tools, which she suggests can be remedied by further research into the sources’ surrounding contexts and digital contact with archivists.3
Ziegler raises concerns about the digital database itself in his 2020 article, reminding us that nothing in the historical digitization process is neutral, from the records themselves to their transcriptions, and that there are always choices made as to what is and is not included in a public database. He also suggests that best practices include digitizing with empathy and consulting cultural experts when working with sensitive materials.4 Additionally, for researchers to use digital databases responsibly, they must recognize that the physical and digital versions of documents are different entities. With this, I emphasize Blaney and Sietring’s argument that you should always cite the format, whether digital or physical, that you actually consulted and drew information from (and I will try to practice what I preach).5 Ultimately, historians must remember that what is digitized is never the whole story of the documents or their context, and should aim to fill gaps and archival silences with further research.
AI and Transcription
After experimenting with different automated transcription software in class, I agree with Cohen’s 2025 assessment that Google Gemini produces the most accurate results.6 The more ethical OCR option, Transkribus, was helpless with the variety of swoopy handwriting samples I gave it, but Gemini was able to decipher documents with slanted and faded cursive. Though I am generally morally opposed to the use of generative AI and LLMs, I have to admit I will likely use Gemini for future large-scale transcription projects due to its accuracy and speed. However, I will always keep Dr. Otis’ advice from Monday in mind: machine transcription is best used as a method of “skimming the document”; you should always review its accuracy before using it as a historical source.
Until next time,
Alexa
- Christof Schoch, “Big? Smart? Clean? Messy? Data in the Humanities,” Journal of Digital Humanities 2, no 3 (2013), https://journalofdigitalhumanities.org/2-3/big-smart-clean-messy-data-in-the-humanities/. ↩︎
- S.L. Ziegler, “Open Data in Cultural Heritage Institutions: Can We Be Better Than Data Brokers?” Digital Humanities Quarterly 14, no 2 (2020), https://doi.org/10.63744/jfyb9zym5am3. ↩︎
- Lara Putnam, “The Transnational and the Text Searchable: Digitized Sources and the Shadows They Cast,” American Historical Review 121, no 2 (April 2016): 377-402, p. 388, 394, https://doi.org/10.1093/ahr/121.2.377. ↩︎
- Ziegler, “Open Data in Cultural Heritage Institutions.” ↩︎
- Jonathan Blaney and Judith Siefring, “A Culture of non-citation: Assessing the digital impact of British History Online and the Early English Books Online Text Creation Partnership,” Digital Humanities Quarterly 11, no 1 (2017), https://doi.org/10.63744/rme2max8ep42.
↩︎ - Dan Cohen, “The Writing Is on the Wall for Handwriting Recognition,” Humane Ingenuity (newsletter), November 25, 2025, https://newsletter.dancohen.org/archive/the-writing-is-on-the-wall-for-handwriting-recognition/. ↩︎
Alexa,
It is enlightening to read about your firsthand experience digitizing BOXES AND BOXES of court records. I agree with your assessment of the need to restrict access to some of this sensitive information, even though these are all public records. It is a consideration that I have wrestled with as well when I took Dr. Robertson’s inaugural Fairfax Court Public History project class last fall. The cases from 1950 that I handled contained intimate details about people’s martial problems that I did not feel needed to be put online even though they are public records and anyone can access them. Oddly enough though I feel that if someone goes through the process of visiting the archive and requesting the records they can look at them. I know it’s a weird dilemma in my brain.
Your work as a historical editor gives you an opportunity to write piece about the choices made and how those choices affect the nature of the digital repository you created. Do you think the omission of a certain subset of these cases calls for a reevaluation of future digitization projects at the courthouse?
Best,
Andrew
Alexa, it was really inspiring hearing about your experiences in digitizing records from the Fairfax County Circuit Court materials. Particularly how these readings aligned with your previous experience but also challenged how you saw your work as a digitizer of a historic collection of records! I would love to talk more about this project with you next time we have class
I think it goes to show that, in this day and age of online databases, which most of us have relied on for our academic research, the unsung heroes of all this are the archivists and others who don’t get the acknowledgment they deserve for digitizing these sources.
It was fascinating to hear about your experience in Dr. Robertson’s course and with the Fairfax Country Circuit Court records. It was also a good reminder that the most valuable sources in an archive are the archivists! As you describe, they determine what’s included and not, and why, and where to find information that may be referenced elsewhere. Inclusion or exclusion could rest on something as arbitrary as not being useful for a certain project, not as a judgment of overall importance or value. And on the use of AI for transcription, I felt the same in class – not wanting to use AI for moral reasons but realizing I would not be able to escape it professionally.
Alexa,
I am pleased you discussed the ethical issues surrounding privacy in digitization. Certain things are clearly defined by law – FERPA, HEPPA, Freedom of Information Act, the Presidential Records Act. However, the vast majority of records do not fall within these frameworks, and leave the ethical decision to the individual repository and archivist. This is an enormous amount of power that collection managers wield. It means that clearly defined policies guiding accessioning, deaccessioning, destruction are necessary, and these terms fully explained to donors and dictated in all deeds of gift. Additionally, archivists should practice empathy when determining if items should be made available or restricted. I tell donors and my employees that our archives will not practice “got-cha” preservation. While we wish to be accurate in what we preserve and digitize, we don’t want to undermine a collection or donor with one-off items.