AI Disclosure

Large Language Model (LLM): Your message is sent to an AI language model provided by OpenAI. It analyzes your message’s content, tone, and intent and generates alternative versions. LLMs generate outputs by identifying patterns in language — they do not understand your situation the way a human would.

Ratonix Learning System: Ratonix records how you interact with suggestions — whether you use, edit, or ignore them, and whether you report outcomes. This system adjusts future recommendations based on what has worked for you and for users with similar communication patterns.

Personalization Layer: Ratonix builds a communication style profile for you and for people you communicate with frequently. This profile influences which tone (Balanced, Softer, or Firmer) is recommended.

Risk Classification: Limited Risk AI System under Regulation (EU) 2024/1689.

Ratonix generates content intended to influence human communication. Under Article 50, users must be informed that content is AI-generated. You are informed of this fact through the Ratonix interface and this disclosure.

Ratonix is not a High-Risk AI System under Annex III because it does not make autonomous decisions about individuals in regulated domains. All suggestions are reviewed by you before use.

• Generate alternative versions of your message in different tones
• Identify potential tone risks in your message
• Recommend communication strategies based on your history
• Learn from your outcomes to improve future recommendations
• Personalize suggestions to your communication style over time

Guarantee outcomes. A suggestion that worked in one context may fail in another. Human communication depends on factors Ratonix cannot observe.

Understand full context. The AI sees the text you enter. It does not know your relationship history, cultural nuances, organizational dynamics, or the events preceding the message.

Replace professional judgment. Do not use Ratonix as a substitute for legal counsel, HR professionals, medical providers, financial advisors, or crisis counselors.

Detect falsehoods. The AI does not fact-check your messages. You are responsible for the truthfulness of your communications.

Ensure cultural appropriateness. AI suggestions may be inappropriate in cross-cultural contexts the model was not trained on.

Operate without bias. AI models reflect patterns in training data, which may include biases. We work to identify and reduce bias but cannot guarantee its absence.

Ratonix suggestions are generated automatically and are not reviewed by a human before delivery.

You are the human in the loop. Every suggestion must be reviewed by you before use. Before sending any Ratonix-assisted message, ask yourself: Is this accurate? Is this appropriate for this relationship? Could this be misinterpreted?

Ratonix improves its AI using interaction and outcome data from users. Full details are in our Privacy Policy. We do not sell this training data. You may opt out of having your identifiable data used for training by contacting privacy@ratonix.ai.

In some contexts you may have legal or professional obligations to disclose AI assistance:

• Legal filings: Many courts prohibit AI-generated content without disclosure and attorney certification
• Academic work: Check your institution’s AI use policy before submitting
• Regulated industries: Financial advisors, healthcare providers, and others may have disclosure obligations
• Political advertising: Some jurisdictions require disclosure of AI-generated political content

Ratonix is not responsible for your compliance with jurisdiction-specific disclosure obligations.

Ratonix uses OpenAI (GPT models) to process messages and generate suggestions. Your message content is transmitted to OpenAI as part of this process under a data processing agreement that restricts how OpenAI may use your data.

If Ratonix generates a harmful, discriminatory, or factually incorrect suggestion, report it to feedback@ratonix.ai. We take all reports seriously.

General: support@ratonix.ai
AI ethics or bias: feedback@ratonix.ai
EU AI Act inquiries: legal@ratonix.ai