Universities Are Drowning in Sustainability Data But Still Can’t Measure It Properly

Universities around the world have spent the past decade branding themselves as sustainability leaders, planting solar panels on rooftops, embedding the Sustainable Development Goals into curricula, and climbing the league tables that rank campuses on their green credentials. Yet a new systematic review of the research published between 2021 and 2025 suggests that the way higher education institutions actually measure sustainability has barely moved into the digital age. The study, published in Discover Education by Yashraj Patil, Harikrishnan Ramachandran, Kainjan Sanghavi and Mahesh Sanghavi, synthesizes sixty peer-reviewed studies and finds a field dominated by static indicators, self-reported data and ranking frameworks, even as campuses generate torrents of real-time operational data that go almost entirely unused for assessment.

The review applies a PRISMA-informed methodology, the transparent screening protocol standard in systematic evidence synthesis. The authors searched the Web of Science Core Collection in January 2026, retrieving 603 records published between January 2021 and December 2025. After title and abstract screening removed 320 records, full-text assessment excluded a further 223, leaving 60 studies for thematic coding. Each study was classified according to its dominant contribution across five domains: sustainability indicators, assessment frameworks, digital and smart campus systems, artificial intelligence and data analytics, and governance and institutional capacity. The authors acknowledge the limitations of relying on a single database and excluding grey literature such as institutional reports, but the resulting dataset offers one of the most complete cross-domain pictures of the field to date.

The first domain, sustainability indicators, is described by the reviewers as a case of expansion without standardisation. Researchers have proposed a proliferation of indicator sets spanning environmental, social, economic and governance dimensions, and these have broadened the conceptual scope of what campuses measure. But the underlying structures remain predefined and static, poorly adapted to the diversity of institutional contexts, and heavily dependent on self-reported data that is often subjective, poorly defined and rarely validated. Because different institutions select, define and weight indicators in incompatible ways, the review concludes that comparability across universities remains weak, undermining the very benchmarking exercise that indicator systems are meant to support.

Assessment frameworks fare only slightly better. Global tools such as UI GreenMetric, the AASHE Sustainability Tracking, Assessment and Rating System, and the Times Higher Education Impact Rankings have undeniably raised awareness and pushed standardisation, giving institutions a common vocabulary for reporting. The synthesis, however, finds that their scoring systems and descriptive reporting structures capture snapshots rather than dynamics, offering little insight into how performance changes over time or responds to interventions. Methodological clarity around weighting and validation is often lacking, and the connection between a framework’s output and an institution’s actual sustainability performance remains obscure. The result, the authors argue, is standardisation without analytical rigor: rankings that are visible and comparable but methodologically thin.

The most striking paradox emerges in the third domain. Campuses have enthusiastically adopted Internet of Things sensors, building management systems, smart meters and digital platforms that continuously monitor energy consumption, environmental conditions and resource use. The reviewed studies include a digital twin framework for monitoring sustainability-related comfort conditions across campus environments, an IoT-enabled smart campus platform supporting real-time operational monitoring, and spatial sustainability dashboards that visualise institutional data for engaged action. These systems dramatically improve data availability and operational visibility. Yet the review finds that digital technologies are used almost exclusively for operational management, not for sustainability assessment. The streams of sensor data rarely feed into the indicator systems or frameworks by which institutions judge and report their performance, because there are no standard architectures for converting operational data into sustainability metrics.

Artificial intelligence shows the highest analytical potential and the lowest adoption. Machine learning, optimisation and predictive modelling studies reviewed in the paper demonstrate capabilities that traditional assessment cannot match: forecasting future performance, detecting anomalies in resource consumption, and optimising energy and occupancy. One study examined the relationship between AI adoption and sustainability ratings in Middle Eastern universities; another applied machine-learning techniques to occupancy prediction for better resource utilisation; a third combined machine learning with genetic algorithms for environmental performance assessment. These are technically sophisticated contributions, but they remain fragmented, typically targeting narrow slices such as energy efficiency rather than holistic, multidimensional sustainability evaluation. The review characterises the field as high potential with limited integration, noting that AI results are seldom connected back to assessment frameworks.

Governance emerges as the binding constraint on all of this. The reviewed studies repeatedly identify fragmented institutional responsibility for sustainability, weak data governance, and limited organisational readiness as barriers to integrating digital data into evaluation. Practical examples illustrate both progress and difficulty: institutional greenhouse-gas accounting frameworks aligned with ISO 14064-1 support reporting and mitigation planning, while other work examines organisational capacity factors and the role of multi-stakeholder engagement in sustainability-oriented governance. But governance structures, the authors find, have generally not evolved to incorporate real-time data and analytics into decision-making, leaving even willing institutions without the institutional machinery to act on the information their sensors collect.

When the five domains are mapped against one another, the structural picture is one of two loosely connected clusters. Indicators, frameworks and governance form the traditional assessment cluster, with strong internal relationships; digital systems and AI analytics form the data-driven cluster, also strongly linked internally. Between the clusters, the connections are weak or absent. The review identifies three specific gaps: an indicator-data gap, in which static self-reported data dominates despite abundant real-time alternatives; a framework-analytics gap, in which assessment schemes lack the predictive and analytical methods that could make them dynamic; and a technology-governance gap, in which digital and AI tools lack the institutional processes needed for large-scale deployment. Quantitative coding of the sixty studies reinforces the pattern, showing governance and indicator research dominating while AI and digital systems remain underrepresented, and empirical approaches outnumbering computational modelling.

The temporal analysis does reveal movement. Research attention to AI-enabled and digital approaches has grown steadily across the 2021 to 2025 window, and the authors detect a shift from static, descriptive evaluation toward dynamic, predictive methods. They argue that the future lies in integrated socio-technical assessment ecosystems in which indicator systems evolve from static checklists into context-aware, continuously updated representations of performance, frameworks incorporate interoperable data architectures for ongoing rather than periodic evaluation, and governance structures embed analytical insight into strategic planning. Aligning such systems with the Sustainable Development Goals and with Environmental, Social and Governance reporting standards, they suggest, would connect campus assessment directly to the decision-making and disclosure processes that increasingly shape institutional accountability.

The review’s conclusions are a sobering read for anyone who assumes that smart campuses automatically produce smart sustainability measurement. The technology exists, the data flows, and the analytical methods have been demonstrated, but the connective tissue between them is missing. The authors call for future research to develop standardised assessment frameworks that integrate digital data infrastructures, AI-enabled analytical tools and governance mechanisms into unified evaluation approaches, validated across diverse institutional contexts to improve reliability and comparability. Until that integration happens, they warn, universities will continue to report sustainability in snapshots while their own buildings stream the evidence needed to tell the full story in real time.

Subject of Research: Sustainability assessment methods in higher education institutions using indicator frameworks, digital technologies and AI

Article Title: Recent advances in sustainability assessment in higher education institutions through indicators frameworks and digital and AI technologies

Article References: Patil, Y., Ramachandran, H., Sanghavi, K., & Sanghavi, M. (2026). Recent advances in sustainability assessment in higher education institutions through indicators frameworks and digital and AI technologies. Discover Education, 5(1), Article 813. https://doi.org/10.1007/s44217-026-02052-1

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02052-1

Keywords: sustainability assessment, higher education, sustainability indicators, smart campus, artificial intelligence, IoT, ESG reporting, Sustainable Development Goals, governance, systematic review, digital twins, data analytics