AI Transparency

EU AI Act 2024/1689 Article 50 disclosure: You are interacting with an AI system. Kaiku uses generative AI models to produce responses, insights and communication analyses. This page clearly describes which AI services we use, for what purpose, and how your rights are protected. This page complements our Privacy Policy (section 5) and Terms of Service (section 7).

1. AI service providers we use

We use the following AI providers. GDPR Art. 28 data processing agreements and SCCs (Standard Contractual Clauses) for EU–US transfers are in place with each of these providers — either as a standalone contract or as part of their standard terms (Anthropic ToS, OpenAI Business Terms, Groq); the current agreement status is available from team@kaiku.live: (a) **Anthropic Claude** (Sonnet, Haiku, Vision) — conversations, analyses, image understanding. API retention 30 days, then deletion. (b) **OpenAI** — Whisper (speech-to-text), TTS (text-to-speech), Realtime API (real-time conversation). Retention 30 days. (c) **Groq Whisper** — fast speech recognition for the chat voice button and Quick-Log. No persistent storage. All providers are contractually bound to Kaiku's instructions and may not use your data for any other purpose.

2. Your data is NOT used to train AI models

None of our AI providers may use your data (conversations, voice recordings, images, health data) to train their own models. This is prohibited in the public Enterprise/API terms of each provider and is reinforced by the standard terms under which Kaiku uses these providers (Anthropic ToS, OpenAI Business Terms, Groq). A short-term API cache (≤30 days) exists solely for abuse monitoring, after which data is deleted.

3. AI can produce incorrect content (hallucinations)

Generative AI models can produce factually incorrect, misleading or incomplete content ("hallucinations"). This is a known limitation of current technology. **Do not rely on AI responses for medical, legal, financial or safety-critical decisions.** Always verify critical information with a professional or authoritative source. Kaiku is not liable for decisions made solely on the basis of AI responses.

4. Kaiku is NOT a medical device

Kaiku is a wellness app, **not a medical device under EU Medical Device Regulation 2017/745.** The app does not diagnose, treat, alleviate or prevent disease. It does not replace a doctor, therapist, psychiatrist, psychologist, physiotherapist or nutritionist. The self-assessments contained in the app (PHQ-9, GAD-7, ECR-R) are reflection tools — **not diagnoses.** In an acute health situation: emergency number 112, MIELI crisis line 09 2525 0111, Sekasin chat, Nollalinja 080 005 005.

5. No automated decisions (GDPR Art. 22)

Kaiku does not make automated decisions concerning you that have legal or similarly significant effects (GDPR Article 22). The AI suggests, you decide. For example: the AI may suggest you consider talking to a therapist, but acting on it is your decision. The AI may flag a risk in couple mode (communication-pattern analysis), but that is **decision support, not an automated intervention.** You have the right to request human review of any AI-generated response or analysis: team@kaiku.live.

6. Hume AI — coming in a later version (estimated Q4/2026)

**Hume AI and other biometric emotion recognition are NOT used at the 1 June 2026 launch.** We plan to add Hume AI's biometric emotion recognition from voice input in a later version (estimated Q4/2026). Under the EU AI Act this is high-risk emotion recognition AI. When it launches: (a) it will be **opt-in**, off by default; (b) it will **not be used in workplace or educational contexts** (AI Act prohibition); (c) the privacy policy will be updated and we will ask for separate consent; (d) you can disable it at any time.

7. Your rights in AI processing

The EU AI Act and GDPR grant you the following rights: (a) **Know when you are talking to an AI** — disclosed clearly in every relevant screen; (b) **Request human review** of any AI-generated response: team@kaiku.live; (c) **Decline to use AI recommendations** — you can always refuse, ask differently or end the conversation; (d) **Withdraw consent for AI processing** (Settings → Consents); (e) **Receive a reasonable explanation of how the AI arrived at its response** (request: team@kaiku.live); (f) **File a complaint** with a supervisory authority: Finnish Data Protection Ombudsman (tietosuoja@om.fi, www.tietosuoja.fi) or the EU AI Act authority once designated in Finland.