AI Transparency Statement

Date: 28 March 2026

Hungarian original (HU): Download the authoritative PDF (HU)

SURWAY’R uses advanced artificial intelligence (AI) to process research data more effectively. We believe transparency matters: what the AI sees and what it does not.

1. What we use AI for

• Schema suggestions: When building surveys, it helps formulate questions.
• Quantitative analysis: It helps identify relationships (e.g. correlations).
• Text analysis: Categorisation and summarisation of open-ended responses.

2. Data handling protocol (safe processing)

• Response-free context: We never send questionnaire responses to AI models (e.g. GPT-4, Claude). The LLM only sees the survey schema (question names, types, options) and aggregated statistical results computed on our servers in advance (e.g. mean, standard deviation, p-values, contingency-table aggregates). Individual, identifiable response data does not leave the SURWAY’R infrastructure.
• No model training: Our contractual terms ensure that providers (OpenAI, Anthropic) do not use the data we send to train their models.
• Zero retention: Where possible, we use APIs that do not retain data after processing.

3. Limitations and responsibility

• AI-generated analyses are probabilistic. They may contain errors or “hallucinations.”
• Results must always be reviewed by a subject-matter expert. AI analysis does not constitute legal, financial, or medical advice.

4. Human oversight

On the platform, decisions made by AI are not automated in a way that would have direct legal effect on data subjects (e.g. rejection). AI is only an assistive tool for the researcher.