AI & Translation Policy

Translation is powerful because it can connect people who otherwise could not communicate. But it also carries risk: mistranslation, missing context, and cultural misunderstanding. This policy sets expectations, protects users, and helps keep translation safe and reliable.

6chatting may use machine learning, AI models, and language processing techniques to translate content between languages. Translation may involve:

• Detection: identifying the source language and target language.
• Conversion: mapping meaning between languages using learned patterns.
• Post-processing: formatting, punctuation, and readability improvements.
• Context hints: prior messages or conversation structure (where available) to reduce ambiguity.

Translation accuracy depends on many factors, including grammar, dialect, slang, noise (for voice), and how much context is present. Common limitations include:

You are responsible for how you use translations. Before you rely on a translation:

• Ask clarifying questions if something seems unclear or surprising.
• Confirm key facts: numbers, dates, locations, instructions.
• Do not treat translations as professional advice.
• Respect cultural differences and avoid assumptions.

Translation features cannot be used to facilitate harm or policy violations. We treat translated abuse the same as abuse in the original language.

• Threats, harassment, hate, or bullying in any language.
• Instructions for wrongdoing, violence, or exploitation.
• Fraud, phishing, impersonation, or manipulation across languages.
• Sexual exploitation, especially involving minors (zero tolerance).

Translation in high-stakes domains can cause serious harm if incorrect. For these domains:

• Medical: symptoms, prescriptions, diagnosis, emergency care.
• Legal: contracts, immigration, court matters, rights and obligations.
• Financial: payments, bank instructions, investment decisions.

For voice/calling translation, accuracy depends on background noise, accents, speaking pace, and microphone quality. To improve results:

• Speak clearly, at a moderate pace, with complete sentences.
• Reduce background noise where possible.
• Repeat key details (numbers, addresses, dates).
• Confirm important points in writing when possible.

Translation can expose children to content they do not fully understand. Parents/guardians should supervise minors and teach them not to share personal information.

Translation requires processing of content to generate translated output. Data handling depends on product features and settings. We aim to minimize data and protect it with reasonable safeguards.

Language models may reflect biases present in training data. We strive to reduce harmful bias and improve translation quality across regions and dialects.

• We encourage reporting biased, offensive, or inaccurate translations.
• We consider cultural context when assessing harm and intent.
• We continuously improve language support over time.

You are responsible for the content you submit for translation. Do not upload content you do not have the right to use, translate, or share (e.g., copyrighted work, confidential documents) without permission.

Language support may vary by region, platform, and feature. We may add, remove, or modify supported languages or models to improve performance and safety.

Report translations that are incorrect, harmful, biased, or unsafe. Include:

• Source language and target language
• The original text/audio context (if possible)
• What the translation output was
• Why it is incorrect or harmful

Misuse of translation features is treated as misuse of the platform. We may apply enforcement actions under our Acceptable Use and Safety Policy, including feature restrictions, suspension, or termination.

We may update this policy to reflect improvements in translation technology, safety practices, legal requirements, or product changes. Updates are posted here with a revised date.

For questions about translation, AI behavior, or safety concerns, contact: