Fresh ideas from museums around the globe in your inbox each week

Step into Antwerp’s Museum Plantin-Moretus and you are stepping into history. Once the bustling workshop of Christophe Plantin, founded in 1555, and later the home of his descendants, the Moretus family, the buildings carry more than just the scent of ink and paper. They house three centuries of lived experience — diaries, ledgers, letters, and casual jottings — meticulously preserved in the very place they were written.
“These records are extremely valuable for research on book production and the cultural life in the Netherlands in the 16th and 17th centuries,” explained Kristof Selleslach, curator of archives. “That is why the museum is strongly committed to digitise these records and make the contents available to a broader audience and to researchers all over the world.”
The catch? Thousands upon thousands of handwritten pages, penned in multiple languages and in the notoriously tricky scrawl of the early modern period.
The museum’s challenge is clear: unlock the stories locked in handwriting that, for most readers, is almost indecipherable.
“Transcribing them in the old-fashioned way would take countless hours of work,” said project employee Tom Eerkens. “The combined effort of artificial intelligence and volunteering helps us to outsource this task.”
The team turned to Transkribus, a handwritten text recognition platform developed at the University of Innsbruck. Its algorithms provide draft transcriptions — imperfect, but good enough to give volunteers a starting point. As Eerkens put it, “These automatic transcriptions still have a little margin of error. That’s why a human eye is still needed to correct these errors, and this is where our citizen science project kicks in.”
Compare this sustained correction process with Cincinnati Museum Center’s handwriting experiment, where the team encountered different limits in abbreviated specimen records.

The project went live just a year ago, and volunteers quickly answered the call. “I saw a call online that you were looking for volunteers for the museum for transcriptions,” recalled volunteer Anneke Van de Voorde. “I sent you an email, and you replied, and it was okay. It was in March from this year.”
For participants like Paule Hofer, the process was initially daunting. “To start, yes, you have to find out a little bit how it’s working. But once you know it, it’s not that difficult,” she said. The key, she noted, was context: “When it’s in a context, very often you can understand it. A single word out of context is sometimes difficult.”
To support this growing community, the museum set up a Slack workspace where volunteers could ask questions, share screenshots of tricky words, or simply exchange observations. “Transcribing does become a group effort rather than a purely individual task,” said Eerkens. “The participants help and support the museum on their own initiative anywhere and anytime they want.”
For Anneke, that sense of camaraderie is part of the joy: “Then I can ask my colleague volunteers on Slack. I put my question there in a thread and they usually reply.”

Volunteers like Anne-Lise Domen see the AI not as a rival but as a partner. “Sometimes you have handwriting that you think, oh my God, I don’t know what I have to make from it,” she said. “AI didn’t know exactly what to make of it either, but at least it gives you some leeway to start with. Obviously you can’t just plug in what the AI did in your online catalogue, but as a starting point, it’s a good one.”
Her enthusiasm is palpable: “I love the whole idea that you make this beautiful archive public and also online public, not only that people come and visit the museum, but also people from around the world can just go online and see what a wealth of information is available here.”

The Slack conversations are part of the transcription work itself. A difficult word can become more legible when another volunteer recognises the context, compares a reading or looks at the surrounding passage. That support connects the machine’s first attempt to a usable record. The museum’s exhibition research then gives the effort a visible destination: corrected documents can help tell the stories of people who lived and worked at the printing house, while also improving the detail available in its inventory.
A different application of human review is IWM’s volunteer-checked image descriptions.
The archive is a window onto everyday life, and volunteers delight in the details. “I like that it’s all things in Antwerp in the 17th century,” said Anneke. “They ordered a lot of butter. They drank only alcohol, beer, and wine, and they needed passports to go to the countryside. It’s a whole saga with the coachman married to the maid. I think you could make a whole television series, like Game of Thrones.”
The museum has already begun weaving these discoveries into exhibitions. “Our community of about 40 online volunteers has already transcribed several archival records related with the printing workshop and the family estates,” said Selleslach. “We have been using these transcriptions in an upcoming exhibition on the women who worked and lived here in the Golden Compasses.”

The project is far from finished. “We aim to make as much transcriptions as possible of as much archive as possible,” said Eerkens. “But that doesn’t mean we don’t have intermediate results. These transcriptions help us to update the existing inventory and make the archive accessible at an item-level description. Above all, we can conclude that this project brings hidden voices from the past back to life.”
One of those voices, lifted from the archives, reads: “I will always remain, beloved father and mother, your humble son, Christophorus Maria Moretus.”
It is a reminder that behind the centuries-old ink and paper are real people — families, workers, merchants — whose lives can once again be heard thanks to the collaboration between AI, volunteers, and the museum that guards their memory.
–
This article is based on the presentation, “Opening up the archives, or how to transcribe with a little help from AI and online volunteers”, by Kristof Selleslach and Tom Eerkens at the Digital Collections Summit 2025 organised by MuseumNext.
Manuel Charr is a journalist working in the arts and cultural sectors. With a background in marketing, Manuel is drawn to arts organizations which are prepared to try inventive ways to reach new audiences.
Fresh ideas from museums around the globe in your inbox each week