AI Use Notice

This Notice explains how the Educoach platform uses Artificial Intelligence (AI), the benefits and limitations, how we protect your data, and what rights you have.

It is published in the context of the transparency obligations of Regulation (EU) 2024/1689 (EU AI Act) and the General Data Protection Regulation (GDPR).

When you interact with Educoach, you are interacting with an AI system. This is also indicated within the platform with explicit markers.

Educoach is a digital platform supporting educational choices that uses AI to help users:

• explore the study programs of Greek universities
• understand their skills, interests, and goals through questionnaires and conversation
• receive personalized recommendations for educational paths and career prospects

The AI acts as a supportive exploration tool — not as a system that makes decisions for you.

The Service relies on four main AI features:

Conversational Chatbot

• Model provider: Google (Gemini)
• What it does: Understands your questions in Greek (and English) and generates answers grounded in the platform's data.
• Model type: General-purpose Large Language Model (LLM) restricted to platform context via Retrieval-Augmented Generation.

Personalized Recommendation System

• What it does: Matches your questionnaire answers and profile data with the available study programs.
• How it works: Computes a relevance score per program based on interests, skills, preferences, and priorities.

Semantic search and Reranking

• Tools: Semantic embeddings (vector text representation) and Cohere's reranker.
• What it does: Identifies and ranks the most relevant excerpts from the programs database to ground the Chatbot's answers.

Automated profile analysis (Profiling)

• What it does: Synthesizes your answers and interactions into a dynamic profile that improves recommendations.
• Limit: Profiling is strictly supportive — it does not produce legal effects for you.

The platform's AI is fed by:

• Database of study programs of Greek universities (public institutional data, titles, durations, scientific fields, professional rights).
• Labor market data and macroeconomic indicators.
• Your own information: questionnaire answers, profile data, and Chatbot questions.

We do not feed the AI with: special categories of data (health, religion, political beliefs, etc.) or with data from other users.

Like any AI system, the Service has substantial limitations you should know:

• Possibility of errors or inaccuracies (hallucinations): The Chatbot's answers may contain mistakes or present as facts information that is not fully accurate. Always verify critical information against official sources.
• Data currency: Data on study programs, labor market, and admission requirements change over time. Despite our update efforts, data may not always be fully current.
• Potential biases: General AI models may reflect biases present in their training data. We work to mitigate them, but cannot eliminate them fully.
• Probabilistic nature: Outputs and recommendations are probabilistic. The same question may return slightly different answers across sessions.
• Dependence on input data: The quality of recommendations depends directly on the accuracy and sincerity of the information you provide.

AI answers do not constitute: professional advice, binding guidance, guarantee of admission, guarantee of professional placement, or legal/financial advice.

For full transparency, we emphasize that the Service does not:

• make solely automated decisions producing legal effects on you (Article 22 GDPR)
• recognize or evaluate emotions, psychological state, or biometric data
• categorize you based on sensitive attributes (race, religion, etc.)
• get used by employers, educational institutions, or other third parties to evaluate or select you
• give third parties access to your conversation data for commercial purposes
• fall into high-risk or prohibited practice categories under the EU AI Act

• We do not use your conversation data to train or improve our own models.
• We do not allow AI providers (Google, Cohere) to use your data to train their general models. This is contractually enforced via the relevant enterprise/API terms.
• Transfers outside the EEA (mainly to the US) are made under the European Commission's Standard Contractual Clauses (SCCs). See the Privacy Policy for details.
• Human oversight: The Provider monitors the system's operation, reviews issue reports, and may intervene in cases of errors or malfunction.

You have the right:

• To know that you are interacting with an AI system — which is indicated within the platform.
• To receive an explanation for the recommendations produced (why a program was suggested): you can ask the Chatbot for clarifications or contact us by email.
• To request human review of any recommendation or answer that concerns you, by contacting the contact email.
• To report errors or misleading answers: this helps us improve the service.
• To delete your profile and conversations, as described in the Privacy Policy.
• To object to automated processing (Article 21 GDPR).

• AI models: Gemini (Google), Cohere Rerank. The list may change; you will be informed of updates.
• The Chatbot does not have real-time internet access. Answers are grounded in the platform's data.
• The platform's source code is not public. However, we describe here the main features for full transparency.

This Notice may be updated when:

• AI features are added, changed, or removed
• model providers change
• new legal obligations arise (e.g. new EU AI Act provisions)

For substantive changes, we will notify you by email or with a prominent in-platform announcement.

For questions, comments, or requests regarding AI use on the platform:

Email: hello@educoach.gr

For complaints regarding the processing of data by AI systems, you also have the right to lodge a complaint with the Hellenic Data Protection Authority (HDPA) — www.dpa.gr.