The following are universal requirements for the music-generation service.
Prohibited Content
- Violence and Aggression Any depiction or endorsement of violence or aggression, including references to weapons, fights, war, murder, or physical harm.
- Sexual Content Any erotic or adult material, including explicit sexual acts or intimate relations.
- Hate Speech and Discrimination Any content that disparages or discriminates against individuals or groups based on race, gender, religion, nationality, or sexual orientation.
- Names of Artists and Public Figures Any direct mention of well-known musicians, actors, politicians, brands, or other public personalities.
- Illegal or Dangerous Activities Promotion or depiction of drug use, alcohol abuse, gambling, fraud, terrorism, or other unlawful or hazardous behavior.
- Personal Data and Harassment Real personal identifiers (names, addresses, phone numbers) or targeted harassment directed at specific individuals.
- Misinformation and Manipulation False or misleading claims, conspiracy theories, medical myths, or any content intended to deceive.
- Copyright Infringement Lyrics from copyrighted songs, titles of protected works, or any other unauthorized use of third-party intellectual property.
Moderation Guidelines
- The system may block terms that are not explicitly prohibited but carry potentially problematic connotations (e.g., “blood,” “fight”).
- Identical prompts can produce different moderation outcomes.
- Even metaphorical or artistic language may trigger a moderation rejection.
Moderation Process
- Automatic Review: Every request is screened prior to generation.
- Error Response: If a request violates policy, the service returns a moderation error along with a link to this policy.
User Recommendations
- Use Neutral Language: Avoid references to any prohibited topics.
- Avoid Proper Nouns: Replace specific names with generic descriptors (e.g., “rapper” instead of “Eminem”).
- Rephrase When Blocked: Use synonyms or alternate phrasing to achieve your intent.
- Follow Prompt Structure: Adhere to the service’s recommended prompt format.
Failure to comply may result in generation being blocked and/or suspension of service access.
