Open Data and AI in Educational Research: Transforming Data into Knowledge
Every interaction a student has on a digital platform generates data. But who controls this information? In this in-depth piece from the MindTech column, we exp
Every time a student accesses an e-learning platform, watches an instructional video, or fills out an online test, they leave behind a digital trail. Until recently, these data fragments remained isolated on the private servers of large tech corporations or were simply deleted.
Today, in 2026, we are witnessing an unprecedented convergence between the Open Data movement and Artificial Intelligence (AI). When educational data is made public and anonymized, and fed to Learning Analytics algorithms, it no longer generates simple statistics, but deep knowledge. It tells us not only if a student is failing, but why they are failing, allowing pedagogy to be redesigned in real time.
In this in-depth analysis for the MindTech column, we will explore how researchers, institutions, and European governments (including Italy) are using the AI-Open Data pairing to create tailored schools, while facing enormous challenges related to privacy, ethics, and the true democratization of access to knowledge.
1. From Numbers to Synthesis: Platforms and Learning Analytics
Contemporary educational research is hungry for data, but constantly clashes with the walls of privacy (e.g., FERPA in the USA or GDPR in Europe). The technological solution to this impasse lies at the intersection of synthetic data and open-source platforms.
A virtuous example is the Education Research Data Platform by Open Education AI. This non-profit platform securely collects vast amounts of school data, but instead of distributing real data on minors, it uses machine learning to generate synthetic data. This "artificial" data maintains the exact statistical properties of the real data (allowing researchers to train their AI models), while simultaneously guaranteeing 100% student anonymity.
The impact of this availability of open data is documented in a solid study published on arXiv entitled Open Datasets in Learning Analytics. Analyzing trends over the last decade (2015-2024), researchers have shown that over 50% of datasets are now open source. This transparency is crucial: it allows the global academic community to replicate experiments (reproducibility), expose the failures of proprietary educational algorithms, and improve understanding of how different human minds learn in digital environments.
Access to data is the first step towards equitable education. We explored how this transparency prevents cultural monopolies in our focus on Open Source Educational Algorithms: Democratizing Digital Learning.
2. Measuring the Impact of AI on Learning
The entry of Generative AI (like ChatGPT or Claude) into schools has often been chaotic. Governments need open data and measurement tools to understand whether this technology is truly helping students or simply making them "lazy."
A pioneer in this regard is the recently launched suite by OpenAI for understanding AI and learning outcomes. This analytical suite is used for collaborations on a national scale (an excellent case study is Estonia, which is testing it on 30,000 students). By collecting aggregated data on how students formulate their requests (prompts) to AI and cross-referencing it with their grades, the platform allows distinguishing between "passive use" (having homework done for you) and "active Socratic use" (using AI as a tutor to explain difficult concepts).
The importance of publicly sharing these results is also supported by the European Commission through the Open Research Europe platform, which has dedicated a specific collection to Education and AI. Making research on Large Language Models (LLMs) in schools open access provides policy-makers with the necessary evidence to legislate, avoiding decisions based solely on tech company marketing.
3. Italy and Europe: Data Literacy and Digital Rights
In Italy, the debate on Open Data and Artificial Intelligence is deeply anchored in the humanistic and ethical component of education. The goal is not to replace the teacher with the algorithm, but to equip students and teachers with the necessary awareness to master data.
The CNR AI-DL Project
A fundamental step in this direction is the project led by the Institute for Educational Technology of the National Research Council (CNR-ITD): AI-DL: Data Literacy in the Age of AI for Education. The project addresses the core of the problem: Data Literacy. If we want AI to transform data into knowledge, we must first teach teachers and students how to "read" data and algorithms (always in full compliance with GDPR). The CNR framework aims to train citizens capable of critically understanding the hidden dynamics of Generative AI.
Open Educational Resources (OER) and AI
On the operational front, the Italian portal Sapere Digitale has initiated a deep reflection on open source AI and the future of education. The article underscores the vital importance of linking the development of educational AI to the world of OER (Open Educational Resources). If the code and training models are closed (proprietary), schools lose control over pedagogical processes. Promoting free software for AI in schools means ensuring that education remains a common good, verifiable and adaptable to the local cultural context.
This availability of pedagogical data allows for the creation of unique learning paths for each student. Discover how in our in-depth analysis: Personalized Learning with AI: Tailor-Made School.
FAQ: Open Data and AI in Educational Research
1. What are "Learning Analytics"? It is the measurement, collection, analysis, and reporting of data about learners and their contexts. It is used to understand and optimize learning and the environments in which it occurs. For example, by analyzing how long a student spends on a paragraph before getting a quiz wrong, AI can deduce which specific cognitive concept they are misunderstanding.
2. What is "Synthetic Data" as mentioned by Open Education AI? Real educational data is extremely sensitive (grades, response times, errors of a minor). "Synthetic Data" is created by Artificial Intelligence: it perfectly mimics the statistics, variance, and mathematical patterns of the original real data, but is fictitious. This allows researchers to test their hypotheses about learning without ever exposing the sensitive data of a single student.
3. Why is it so important for educational datasets to be "Open"? If data on the effectiveness of an educational AI is locked in the servers of the company that produced it (Black Box), no independent scientist can verify their claims. Open datasets allow for scientific reproducibility: the academic community can test algorithms to verify if they promote learning or if, for example, they discriminate against non-native speaking students.
4. What is the difference between "Computer Literacy" and "Data Literacy"? Computer Literacy is knowing how to use a computer (e.g., writing a Word file or using a browser). Data Literacy (promoted by projects like the CNR's) is the ability to read, work with, analyze, and argue with data. In the age of AI, it means understanding how a machine draws conclusions from your inputs and recognizing potential logical errors or algorithmic biases.
5. Does the use of AI in schools violate European GDPR? It depends on how it is implemented. The use of commercial AI platforms (which transfer student data to America to train their models) is often in violation of GDPR. For this reason, European institutions are pushing for the use of open source models installed on local (or national) servers where data is anonymized, ensuring that students' digital traces never become a commercial product.
Conclusions: Guardians of the Digital Future
Artificial Intelligence is not magic; it is a formidable statistical engine fueled by tons of data. The way we choose to collect, share, and interpret the educational data of our children will determine the shape of the school of the future.
If this data remains locked in a monopoly regime, we will witness the predatory commercialization of the education system. If, on the contrary, as European and Italian projects are demonstrating, we push for the convergence of Open Data, Open Source, and Artificial Intelligence, we will have, for the first time in history, an accurate map of how the human mind learns.
This transparency is the only tool capable of transforming the cold extraction of metrics and percentages into true pedagogical knowledge: a shared understanding that leaves no student behind, while simultaneously protecting their inviolable emotional and cognitive privacy.
Bibliographic References and Sources
To ensure scientific accuracy and technological currency, this article drew upon the following primary sources:
- Open Platforms and Research on Data and Learning Analytics:
- Impact of GenAI on Learning:
- OpenAI – Understanding AI and learning outcomes (Measurement suite and national collaborations). Link
- Italian Context: Data Literacy and Open Educational Resources: