evidoria

← Back to browse

Good practice Imported

Bulgaria's 41-Page National AI-in-Schools Guidelines — One of Europe's First

Bulgaria · Sofia · See the Bulgaria profile · See the Sofia profile

Evidence: Descriptive / self-reported Top 58% 47/100 · Ask Evidence Copilot about this practice

Bulgaria's Ministry of Education issued a 41-page voluntary national guide on AI use in schools in February 2024, among the first in Europe, then proposed a 2025 draft strategy adding risk classification and mandatory human review for high-stakes uses like grading.

Bulgaria's 41-Page National AI-in-Schools Guidelines — One of Europe's First

Details

Promoter
Bulgarian Ministry of Education and Science
Period
19 February 2024 (guidelines); draft risk-tiered strategy proposed May 2025
Keywords
AI policy, teacher guidelines, K-12 education, academic integrity, risk classification

Context

On 19 February 2024 Bulgaria's Ministry of Education and Science published a 41-page national guidance document on the use of artificial intelligence in schools, aimed at teachers, school officials, parents and students. Then-Education Minister Galin Tzokov described Bulgaria as among the first countries to introduce such guidelines for schools.

Objectives

The guidance explains what AI is and how chatbot-style tools work, and states explicitly that 'the use of AI in education must be constantly monitored and evaluated for its impact, taking into account the pace of development of technology,' stressing that AI should support, not replace, the decision-making authority of students, teachers and administrators.

Activities

The document sets out example uses such as lesson creation and identifying students at risk of dropping out through analysis of grades and absences. It is not mandatory or binding — adoption is left to individual schools. A follow-up Draft Strategy for the Development and Integration of Artificial Intelligence in Bulgarian Education, proposed in 2025, goes further: it proposes classifying educational AI uses by risk level (aligned with the EU AI Act paradigm), mandatory human-in-the-loop review for high-stakes decisions such as grading and admissions, mandatory Data Protection Impact Assessments, an inter-ministerial governance body, and a national registry of approved AI tools.

Conclusions

As of this review, the 2025 draft strategy remains a proposal rather than adopted policy, so its governance provisions should be read as a stated direction, not yet an implemented safeguard. The 2024 guidance itself is voluntary, with no mandate or enforcement mechanism requiring schools to follow it.

Implementation

Indicative cost
Low (< €50k) — Cost is limited to policy development and drafting by the Ministry of Education and Science; no implementation funding, training programme or enforcement budget is documented for the 2024 guidance, and the 2025 strategy's registry/DPIA infrastructure costs are not yet specified since it remains a draft.
Time to results
Short (< 1 year) — Guidelines published 19 February 2024; a more binding draft strategy was proposed in 2025 but had not been adopted as of this review.
Staffing & skills
Bulgarian Ministry of Education and Science policy staff, proposed inter-ministerial governance body (2025 draft strategy, not yet established)

Conditions for success

  • voluntary buy-in from individual schools, since the guidance is not mandatory
  • alignment of the future risk-classification system with the EU AI Act framework
  • capacity to run Data Protection Impact Assessments and maintain a national registry of approved AI tools once the 2025 strategy is adopted

Common failure modes

  • non-binding status means impact depends entirely on voluntary school-level adoption, which is not tracked or reported
  • the more substantive safeguards (risk classification, mandatory human review, DPIAs) exist only as a 2025 draft, not yet adopted policy

Where it fits

Governance type
national ministry guidance (soft law), voluntary
Scale
national
Income level
high-income, EU member state (Bulgaria)

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful