What happened
Newsbytes.PH reports that a University of the Philippines policy brief proposes using artificial intelligence and machine learning to monitor the country’s public libraries, identify facilities at risk of closure and guide government funding. The researchers analyzed 2025 National Library Directory data with four classification models; Gradient Boosting performed best in the reported tests. The brief recommends an AI-enabled monitoring system, a digital-readiness index and predictive funding models. The study’s findings and proposals have not been independently confirmed from the source provided.
Newsbytes.PH reports that researchers at the University of the Philippines Center for Integrative and Development Studies’ Program on Data Science for Public Policy published a policy brief proposing AI and machine learning as tools for redesigning the country’s public-library system. The report says the brief calls for changes to Republic Act No. 7743, the 1994 law requiring congressional, city and municipal libraries and barangay reading centers. The central issue identified by the researchers is not simply where libraries were created, but whether they remain operational and capable of serving their communities.
According to Newsbytes.PH, the researchers cited the 2025 National Library Directory and said only about 30% of registered libraries were currently open. The article attributes that figure, along with concerns about inadequate resources, limited technology skills, weak IT services and poor internet connectivity, to the policy brief. It also reports regional differences: the National Capital Region and Region I were described as relatively stable, while the Bangsamoro Autonomous Region in Muslim Mindanao, Caraga and Region II were identified as having low operational capacity and a higher risk of closure.
Newsbytes.PH says the researchers tested four machine-learning classification approaches: Random Forest, Logistic Regression, a Support Vector Machine using an RBF kernel and Gradient Boosting. Gradient Boosting reportedly produced the strongest results, with accuracy and weighted recall of 0.788 and a weighted F1 score of 0.736. Random Forest reportedly reached 0.727 accuracy, while Logistic Regression reached 0.667. The article says the SVM performed poorly because of severe class imbalance in the dataset. These are results reported by the outlet from the policy brief; the source provided does not include the dataset, validation design, confidence intervals or an independent replication.
The brief’s main proposal, as described by Newsbytes.PH, is a National AI-Enabled Library Monitoring System, or NAELMS, that would classify libraries in real time as active, at risk or closed. Other proposals include a Library Digital Readiness Index measuring infrastructure, connectivity and learning engagement; annual AI-assisted assessments; a national data dashboard; predictive maintenance; and AI resource centers attached to established provincial libraries. The article also reports a proposal to convert nonperforming libraries and reading centers into hybrid community learning centers offering e-learning, digital-skills training and access to government services.
Source details: newsbytes.ph ↗
Why it matters
The proposal would shift library policy from counting whether facilities were established to tracking whether they remain open, connected and useful. If implemented responsibly, data-driven monitoring could help officials identify underserved communities and prioritize scarce funds. But the reported model performance is below perfect accuracy, and the source does not establish that the system has been deployed, independently validated or shown to improve library outcomes.
The reported proposal addresses a practical weakness in public-service administration: a legal requirement to establish facilities may not reveal whether those facilities have staff, connectivity, equipment or regular users. A monitoring system could give officials a more current picture of library conditions than a static directory. In principle, that may help distinguish a library that exists on paper from one that can actually provide reading materials, internet access, digital training and public services.
The funding proposal is potentially consequential because it would change how government resources are allocated. Newsbytes.PH reports that the researchers favor directing funds toward libraries identified as at risk instead of relying on uniform allocations regardless of local conditions. Such targeting could benefit rural and municipal libraries that lack infrastructure, but it could also create new inequities if the underlying data are incomplete or if facilities with weak records are treated as low priority rather than as places needing immediate support.
The model results illustrate both the promise and the limits of the idea. A reported accuracy of 0.788 means the tested approach did not classify every case correctly, and the weighted metrics may conceal weaker performance for smaller or underrepresented classes. The article specifically notes severe class imbalance affecting the SVM, but it does not say how imbalance affected the other models or whether the system was tested separately across regions. A classification error could have practical consequences if it influences inspection schedules, funding decisions or a library’s public status.
The policy brief reportedly acknowledges privacy, cost, infrastructure and AI-literacy barriers. Those limitations are central rather than incidental. Library monitoring could involve information about facilities, staff, connectivity, usage or community engagement, and the source does not explain what personal data would be collected or how long it would be retained. Nor does it show that AI would outperform trained human assessors in real operating conditions. The public value therefore depends on transparent data practices, human review and funding decisions that do not treat a model’s output as an unquestionable verdict.
What to watch next
Key questions are whether the Philippine government adopts any of the proposed systems, what data would be collected and how libraries would challenge incorrect classifications. Further scrutiny is also needed on regional representation, class imbalance, privacy safeguards, infrastructure costs and whether predictive funding would supplement rather than replace baseline support. The source does not provide a timetable, confirmed government commitment, implementation budget or independent evaluation.
The first development to watch is whether any Philippine national or local government agency formally adopts the proposed system. Newsbytes.PH reports recommendations, not a confirmed deployment, procurement decision or implementation schedule. The source also does not identify a committed budget, responsible agency, technical vendor or timetable for amending Republic Act No. 7743. Until those details emerge, NAELMS and the related dashboard remain policy proposals rather than operating public infrastructure.
Independent scrutiny should focus on the data and evaluation methods. The source says the researchers used the 2025 National Library Directory, but it does not state how many libraries were included, how labels such as active, inactive or closed were established, how the data were divided for testing, or whether performance was measured on geographically separate samples. Reviewers should also examine whether the reported weighted scores mask poor results for smaller regions or libraries with limited records.
Privacy and accountability rules will be important if monitoring expands beyond facility-level information. The brief reportedly recognizes privacy concerns, but Newsbytes.PH does not describe consent rules, data minimization, access controls, appeal procedures or safeguards against using proxy variables that disadvantage poorer communities. A workable system would need a clear process for correcting inaccurate records and a human decision-maker who can override a classification when local evidence conflicts with the model.
The source leaves unresolved whether AI-assisted targeting would add resources to struggling libraries or merely make it easier to withdraw support from facilities labeled nonperforming. Further reporting should establish how the proposed Library Digital Readiness Index would be calculated, whether communities would participate in assessments and how funding outcomes would be audited. It should also test whether hybrid learning centers preserve core library functions, and whether improved connectivity and staff training—not prediction alone—produce measurable gains in access and digital literacy.