The system provides students with fast, personalized, and reliable support in large courses, while freeing teachers from repetitive inquiries and enabling consistent application of course policies. AIVA represents a new generation of educational AI systems – not as a replacement for teachers, but as a transparent and controlled advisory layer that scales pedagogy in large educational systems.
Unlike generic chatbots that generate free-text responses and often provide uncertain or inaccurate information, AIVA is a pedagogically constrained advisor based on the theory of Self-Regulated Learning (SRL). The system does not replace the teacher; instead, it scales and consistently implements their support, helping students plan their obligations, monitor their progress, and make informed academic decisions.
The concept is inspired by Zimmerman’s SRL model, which distinguishes three key phases of learning: planning, performance monitoring, and reflection. Through interaction with AIVA, students gain a clearer understanding of course structure, minimum and desired passing thresholds, deadlines, and grading criteria. Integration with the Moodle gradebook enables personalized advice on current status, risk alerts, and clarity in academic decision-making.
A key advantage of the system is its transparency and control: AIVA uses exclusively course data and rules defined by the teacher, clearly indicates the sources of its responses, and is fully compliant with GDPR. This ensures pedagogical coherence and prevents the replacement of the teacher’s role.
AIVA is currently implemented in more than 20 courses at FOI, covering approximately 1,000 students across various study programs, including undergraduate and graduate levels. Upon first use, each student accepts the Terms of Service, further ensuring ethical standards and privacy protection.
Prof. Igor Balaban, PhD, team leader, emphasizes:
“Through the development of AIVA, our students receive fast, accurate, and scalable support, while teachers can concentrate their time and energy on creative and pedagogically more demanding aspects of teaching. This is an example of how technology enhances the quality of education and supports self-regulated learning.”
In addition to Prof. Balaban, AIVA was developed by Miran Zlatović, PhD, and assistant Marko Matus, who designed the concept and system functionalities, while GTN Solutions, as a Moodle partner, technically implemented the solution. In cooperation with international partners, FOI plans further development of AIVA through Erasmus+ projects, including standardization of the pedagogical AI framework, piloting in large courses, and evaluation of its impact on self-regulated learning and the reduction of teachers’ administrative workload.
The next development steps include proactive alerts about deadlines and risks of not meeting requirements, stronger Moodle integration, additional levels of response explainability, and flexible AI model selection to optimize costs.
The long-term goal is for AIVA to become a reference model of AI-based advising in higher education – a scalable, pedagogically grounded system that supports students across Europe.
