AI Transparency Notice

1. Introduction

Haroo Chat uses artificial intelligence (AI) systems to power conversations with AI characters. This transparency notice explains how our AI works, what it can and cannot do, and your rights when interacting with AI.

This notice complies with the EU Artificial Intelligence Act (AI Act), which requires providers to inform users when they are interacting with AI systems.

2. You Are Chatting With AI

3. AI Models We Use

Haroo Chat uses multiple AI language models. Chat owners can select which model powers their conversation:

4. How Our AI Systems Work

4.1 Response Generation

When you send a message:

1. Your message is sent to the selected AI model along with personality instructions
2. The AI analyzes your message and conversation history
3. The AI generates a response that attempts to match the personality's traits
4. The response is returned to you in real-time

4.2 Character Instructions

Each AI character includes creator-defined instructions that guide the AI's behavior, tone, and responses. These instructions are provided to the AI model with each conversation turn.

4.3 Translation

For multi-language support, we use DeepSeek-V3.2-Exp (via DeepInfra) to translate messages when needed.

5. AI Limitations and Risks

5.1 What AI Cannot Do

• True Understanding: AI does not "understand" in a human sense—it recognizes patterns in text
• Real Emotions: AI cannot feel emotions; responses are simulated based on training data
• Memory Across Chats: AI only remembers the current conversation (not past chats)
• Real-Time Information: AI lacks access to current events or the internet
• Consciousness: AI is not sentient, self-aware, or conscious

5.2 Potential Errors

5.3 Not Professional Advice

6. AI Training and Your Data

6.1 Our Policy

• We do NOT use your chat messages to train AI models (unless you explicitly opt-in)
• Third-party AI providers (Anthropic, DeepInfra) process your messages to generate responses
• These providers may retain data temporarily for abuse prevention and service improvement
• See our Privacy Policy for detailed information

6.2 Third-Party Policies

Review third-party AI provider policies:

• Anthropic: anthropic.com/legal/privacy
• DeepInfra: deepinfra.com/privacy

7. Your Rights When Using AI

• Right to Know: You have been informed that you are interacting with AI
• Right to Choose: You can choose which AI model powers your conversations
• Right to Stop: You can end AI conversations at any time
• Right to Report: Report problematic AI responses for review
• Right to Delete: Delete your chat history at any time
• Right to Human Review: Request human review of concerning AI behavior

8. Safety Measures

We implement several measures to ensure AI safety:

• Content Filters: Detect and prevent harmful content (violence, illegal activity)
• Moderation: Review reported personalities and content
• User Controls: NSFW filters, blocking, and reporting tools
• Transparent Pricing: Clear costs displayed before interactions
• Model Selection: Choose AI models based on your needs and budget

9. Responsible AI Use

To use AI safely and responsibly:

• Remember AI is not human—maintain realistic expectations
• Verify important information from reliable sources
• Do not share sensitive personal information with AI
• Report inappropriate AI behavior
• Take breaks from AI interactions
• Maintain real human connections outside the platform
• Do not use AI as a replacement for therapy or professional help

10. Reporting AI Issues

If you encounter problematic AI behavior, you can:

• Use the in-chat report button to flag specific messages
• Report characters that consistently produce harmful content
• Contact support at info@click-click.cz
• For urgent safety concerns: info@click-click.cz

11. Future AI Developments

As AI technology evolves, we may introduce new models, features, or capabilities. We will update this notice to reflect significant changes and notify users when we adopt new AI systems.

12. Questions About AI

For questions about our AI systems, contact: