AI and Transparency Notice — FizzUp

1. In brief

FizzUp uses artificial intelligence (AI) to:

• conduct practice conversations,
• generate feedback and suggestions for improving your utterances,
• propose exercises and recommendations,
• help personalize learning.

By using FizzUp, you interact with an AI system and not with a human.

2. Where in FizzUp we use AI

AI may be used, among other things, for:

• conversations (text and/or voice),
• analysis of utterances and generation of suggestions,
• progress summaries,
• extracting learning patterns (e.g. "weak points", "favorite style").

3. Limitations of AI

AI may:

• make mistakes,
• fail to fully understand the context,
• misrecognize speech (depending on audio quality),
• propose phrasings that do not fit the situation.

Treat the results as practice suggestions, not as certainties.

4. What AI does not do

FizzUp is not:

• an official exam or certification,
• a legal, medical, psychological, financial or immigration advice service,
• a tool for making employment decisions,
• a system for biometric identification based on voice.

5. Voice and transcripts

In voice features, we process your speech to enable the Service to work (transcription and feedback).

In Speak (Gemini Live) mode, the audio stream may be transmitted directly to the AI provider from your device, and our backend only issues a token for the connection.

6. Responsible use

We recommend that you:

• do not enter particularly sensitive data into conversations,
• do not paste confidential third-party data unless you are authorized to do so,
• verify key information where it has "high-stakes" significance.

7. Human review

We do not routinely listen to private sessions for the purpose of "manual evaluation". Human access may occur only to a limited extent, where it is necessary for:

• handling support requests,
• security and abuse detection,
• debugging and quality assurance,
• meeting legal obligations.

8. Contact

Questions about AI and security: ai@fizzup.ai