Artificial intelligence: The legacy challenges of making heritage collections computationally ready

Alexis Tindall, Manager, Digital Innovation, University of Adelaide, Ingrid Mason, Project Manager, Australia National University, and Sydney Shep, Reader in Book History, The Printer, Wai-te-ata Press, Victoria University of Wellington.

The team speak of the challenges of making collections computationally accessible using AI technologies. While the GLAM community has been digitising content for decades, the new demand for full-text or structured machine-readable data has led to the need for a fresh approach. The hurdles to making collections computationally accessible include turning handwritten documents into full text, making structured data accessible, licensing digitised materials, and making license information machine-readable. The local AI for LAM chapter has delivered community webinars to advance awareness and discussion of how new technologies affect the role of information custodians.

The speakers discuss how various teams have implemented the use of web-based AI services. Projects have broadly fallen into two categories: those that apply machine learning/AI to the challenges of data curation, collection management, and sharing, and those that use machine learning/AI for conducting research or enabling new forms of use of GLAM-sourced data. The challenges and opportunities around sustainability, creativity, and indigeneity in research are also discussed.

Many projects are experimental pilot projects that do not result in long-term investments or embedded business practices. The team urge researchers to consider long-term viability from the outset and work with professionals who can build it into every component of their workflow. They also emphasise the need for creativity - encouraging institutions to provide resources that shape new ways of thinking and doing, and to create space that encourages collaboration with creative coders and digital media artists.

Finally the team raise the need for cultural AI, recognising the importance of unlocking the largest corpus of digitised data, building capability and forging new pathways for community-driven and real-world impactful research.

This is an AI generated summary. There may be inaccuracies

Previous

Co-designing an exhibition with children - can it possibly be child's play?

Next

Adventures in Creating Original Virtual Reality Content with Limited Resources, Time and Budget