Archivists assist users reminiscent of historians, genealogists, students, or residents in locating, accessing, and interpreting archives. Archive reference services have long been considered Services that provide understanding and dialogue between archivists, users and archives to make documentary objects more accessible and usable.
In recent years the next have been introduced: artificial intelligence (AI) in cultural heritage institutions reminiscent of libraries, archives, museums and galleries.
Researchers are studying how AI impacts and can impact archival services from the automation of file management to the organization of archives and latest types of digital archives. There was loads of discussion about it the advantages of AI on the subject of User support.
Among the AI-powered technologies are ChatGPT
can support some facets of archival reference services. However, use requires human supervision.
Using some examples of an actual conversation with it ChatbotIt is feasible to look at the relevance of this AI-supported technology as an archive assistant – and likewise its limitations.
(Shutterstock)
A glossary of key concepts
Suppose a social science student must seek the advice of and examine historical written sources. However, the scholar isn’t sufficiently conversant in the essential principles Archival science – methods to discover, acquire, authenticate, preserve and access Records of lasting value.
For example, after making an internet request for his country's national archives, Library and Archives CanadaThe student was invited to personally view the archives of interest to him, as they aren’t accessible online.
If the scholar wants to make sure a basic orientation for a visit, they might use ChatGPT to assist and ask the query: “What are archives?”
When I asked this query, the reply generated gave suggestions on methods to use this chatbot as a glossary. The answer focuses on the character of archives and their importance in providing authentic evidence of activity: “Archives discuss with a set of historical records preserved for long-term use.” These records will be in various formats including paper documents, Photos, audiovisual materials, digital files and more.”
Chat GPT defined what the word “archives” means and what characteristic properties these authentic documentation objects have. What helped with the definition was that ChatGPT qualified archives more accurately as documentary evidence preserved for evidentiary purposes, whether for administrative, historical, or scientific purposes.
A practical guide or manual
Suppose a student recognizes that in an email he received from his country's National Archives, the archivist used the word “to seek out help” as a search tool. The student wants to know how a finding aid works.
I asked this query to ChatGPT and the reply generated reflects a set of instructions typically provided by archivists to users. It notes that a finding aid helps users change into more conversant in the scope, layout and organization of archives, in addition to any limitations.
![List of text answers to the question “How does a finding aid work for locating archives?”, including collection naming overview, scope and contents, arrangement and organization, series and subseries, container list, index terms and topics, access restrictions.](https://images.theconversation.com/files/562285/original/file-20231128-27-jrgh2u.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip)
(Siham Alaoui), Provided by creator (no reuse)
A search engine or an internet catalog
Or, if a student decides to go to an archives consultation room, they could notice that some archives contain a reference to archival materials from the National Library and Archives of Quebec (Library and National Archives of Quebec).
The student doesn’t live in Quebec but is willing to travel there to view these archives. It is feasible to question ChatGPT – as I did – to see if it will possibly provide him with relevant information in order that he is best prepared for his visit, and ask: “Give me an example of the Bibliothèque’s collections et Archives Nationales du Québec.”
For each of the generated examples, a temporary description of the archives included in each category is provided (photos, maps, texts, audiovisual materials, etc.). However, this question doesn’t provide exact names of the archives. Nevertheless, even this limited query of ChatGPT suggests that it will possibly also improve the visibility of heritage institutions' archival collections and support higher dissemination of cultural heritage.
![List of text answers to the question](https://images.theconversation.com/files/562259/original/file-20231128-25-jbam7p.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip)
(Siham Alaoui), Provided by creator (no reuse)
“Archive information”
Getting essentially the most out of the archival user experience requires what archival researchers Elizabeth Yakel and Deborah A. Torres call a set of translation skills . As my examples show, using a generative AI tool like ChatGPT allows users to develop a few of this archival intelligence autonomously.
However, there are limitations to be noted. In my examples, a lot of the responses generated by ChatGPT contain concepts like “document” and “record” used interchangeably. Considering how this impacts search requires a trained eye conversant in the sources and variations in terminology across time and systems.
Professional familiarity with terminology
When I attempted to unravel this problem by asking one other query: “What is the difference between a record and a document?”, ChatGPT couldn’t properly distinguish the most important difference between the 2.
The answer given was: “Documents are a subset of records. Not all records are documents, as records may contain non-documentary forms of knowledge reminiscent of data entries, logs, or other records or activities.” This answer is wrong because in archival science Records are documents because they’re products of inscription on a comparatively stable medium.
![Text answers listed following the question “What is the difference between a record and a document?”](https://images.theconversation.com/files/562266/original/file-20231128-23-ujnyvr.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip)
(Siham Alaoui), Provided by creator (no reuse)
Archivists are still being sought
Many scholars in archival studies have examined the types of records in digital environments. The heads of those are research groups The International Research on Permanent Authentic Records in Electronic Systems (InterPARES)led by Luciana Duranti, Professor Emeritus of Archival Studies on the University of British Columbia.
InterPARES provides insights into how records that ought to exist in a documentary medium and have lasting content can manifest in a different way in digital environments. So records are documents. This makes it clear to us that ChatGPT may provide false or misinformation. Another problem is that users may misinterpret ChatGPT answers.
Therefore, archivists are still invited to play one Documentary mediation Role with users. This ensures that the latter make good use of AI-powered technologies to enhance their archival intelligence and higher understand archives and their terminology.