AI Safety Notice

Version: 1.0 Effective Date: 20 May 2026

What you should know

HablaYa is a language practice service where your conversation partner is an AI model. AI-generated responses are produced automatically and may contain errors: factual, grammatical, cultural.

AI can make mistakes. Verify important information.

Limitations

• AI does not replace a live teacher, native speaker or certified expert.
• AI does not provide legal, medical, financial or other professional advice. If a response relates to such matters, consult a qualified human.
• AI can produce plausible-sounding but inaccurate facts — especially dates, names, quotes, idiomatic translations.

Providers

AI responses are generated via third-party models:

• OpenAI (GPT-4o-class) — primary model for dialog, helpers (translate / what-to-reply) and speech recognition/synthesis.
• Google (Gemini Flash) — supplementary provider for select functions.

The full list of providers, jurisdictions and cross-border transfers is described in the Privacy Policy.

Voice messages

If you consent to voice processing, your audio messages are sent to the provider for speech-to-text. After transcription, audio is retained for 30 days and then deleted. See Data Retention Schedule for details.

If voice consent is not given or has been revoked, the service continues to operate in text mode.

Complaints and feedback

If AI produces a response you find offensive, harmful or misleading, tap the flag icon next to the message. Complaints help us improve prompts and filters.

For any questions regarding AI processing of your data — contact privacy@hablaya.work.

Change history