Jakarta, INTI - The National Library of Indonesia (Perpusnas) is expanding its use of artificial intelligence (AI) to support the preservation of ancient Indonesian manuscripts while strengthening the nation’s collective cultural memory.
Perpusnas Head E. Aminudin Aziz said the technology is being developed to help identify the level and types of damage found in ancient manuscripts.
“Perpusnas has continuously explored the use of AI in manuscript preservation. One of our staff members, who is an alumnus of the Bandung Institute of Technology (ITB), has developed an algorithm to identify the severity and types of damage affecting manuscripts,” Aminudin said.
AI Supports Manuscript Damage Assessment
The application is expected to help identify the causes of manuscript deterioration, including damage caused by insects, humidity, and fungal growth.
Once the cause of deterioration has been identified, the technology could also assist conservators in determining the appropriate treatment for each manuscript.
AI is additionally being explored as a way to reduce subjectivity in assessing manuscript conditions, as the process is still largely conducted manually.
Manual assessments can produce different results depending on the individual responsible for evaluating the condition of a manuscript.
“Damage assessments have so far been highly manual, meaning that different people may reach different conclusions. AI is expected to provide more consistent results after undergoing repeated testing,” Aminudin said.
Perpusnas expects the technology to deliver more consistent assessments as the AI system undergoes further testing and refinement.
Regional Scripts Remain a Challenge
Despite its potential in supporting preservation efforts, AI technology for reading ancient manuscripts and translating them into Indonesian has yet to reach its full potential.
Automated text recognition is generally easier to apply to widely used writing systems, including Latin, Chinese, and Arabic scripts, as well as several major languages.
However, greater challenges arise when AI is applied to ancient Indonesian manuscripts written in various regional scripts.
Automated recognition of Sundanese, Javanese, and other traditional Indonesian scripts remains less advanced than recognition of more widely used writing systems.
“The challenge lies in manuscripts written in scripts such as Sundanese, Javanese, and other regional scripts, which cannot yet be automatically recognized at the same level of accuracy. Therefore, we plan to strengthen collaboration between manuscript specialists and AI experts to develop technology that can provide significant benefits, particularly for preserving ancient Indonesian manuscripts,” Aminudin said.
Perpusnas plans to strengthen collaboration between manuscript experts and AI specialists to address these technological limitations.
However, developing such technology has not yet become a major priority for Perpusnas due to budget constraints.
If the research can be carried out, AI could potentially improve manuscript readability while enabling more accurate and objective text recognition and condition assessments.
Greater readability and accuracy could also expand access to ancient manuscripts, allowing them to be studied and utilized by both Indonesian and international communities.
Perpusnas Records 261 Manuscripts in Khastara
As of September 2026, Perpusnas recorded 261 ancient Indonesian manuscripts registered in the Khazanah Pustaka Nusantara (Khastara) system.
The registered manuscripts include collections owned by individuals as well as institutions.
Perpusnas is encouraging individuals and communities with ancient manuscripts to actively register their collections through Khastara.
The initiative is expected to make Indonesia’s ancient manuscript collections more accessible and enable them to reach a wider audience.
Conclusion
The adoption of AI by the National Library of Indonesia marks an important step toward modernizing the preservation of the country’s ancient manuscripts. By improving damage detection, assessment consistency, and potentially text recognition, AI could make these valuable cultural resources more accessible and better preserved for future generations. However, challenges involving regional scripts, technological development, and limited funding remain key areas that require further research and collaboration.
Read more: AI Advances Rapidly, Deputy Minister Nezar Says Regulations Must Keep Pace