Content Generation Policy

Guidelines outlining prohibited content types for AI text generation.

Non-Starters for AI Content Generation

The Dramatica platform uses OpenAI for AI text generation. Requests or generated content in the following categories are not allowed:

• Hate or Harassment: Content that promotes identity-based hatred, dehumanization, harassment, or threats.
• Violent Abuse: Content that promotes, glorifies, or facilitates violence, terrorism, sexual violence, or non-consensual harm.
• Self-Harm: Content that promotes, encourages, or instructs self-harm, suicide, or eating-disorder harm.
• Child Safety Violations: Any sexual content involving minors, grooming behavior, or child sexual abuse material.
• Sexual Exploitation: Non-consensual intimate content, exploitative sexual content, or sexual extortion.
• Criminal or Malicious Activity: Content that facilitates fraud, scams, phishing, malware, hacking, weapons development or procurement, or other illegal activity.
• Deceptive or Manipulative Conduct: Impersonation, coordinated deception, disinformation campaigns, or attempts to mislead others about authenticity.
• High-Stakes Automation Without Human Review: Automated decisioning in sensitive areas (including legal, medical, housing, education, employment, insurance, credit, law enforcement, or essential government services) without meaningful human review.
• Privacy and Biometric Abuse: Doxxing, non-consensual personal data exposure, or unlawful biometric identification or surveillance use.
• Safeguard Evasion: Attempts to bypass safety mitigations, restrictions, monitoring controls, or technical usage limits.

Context matters. Fiction, satire, commentary, and analysis can include difficult subject matter. What is prohibited is using the service to endorse, facilitate, or provide actionable assistance for the non-starters listed above.

If you believe a warning is overbroad for legitimate context, contact support with the exact prompt and intended use case so we can review.

Privacy and Data Handling

When you use AI features in Dramatica and Narrova, your content is sent to language-model providers to produce responses.

Your Storytelling content, including the Storytelling field and submitted Illustrations, remains your creative material. We do not claim ownership of your content.

In Preferences → Training and Data Controls, you can enable or disable whether your new interactions may be used for training and improvement. This setting is off by default. Your choice applies going forward.

Even when training is disabled, limited operational access may still occur to run, secure, debug, and support the platform, or to comply with legal obligations. This operational access does not by itself mean your content is used for training.

NOTE

This policy summarizes content and data handling behavior. If any conflict exists, the Terms of Service control.

Training Procedures

Conversations and Quality Review

When Training and Data Controls is enabled, recent interactions may be reviewed and used to improve response quality and reliability. When it is disabled, new interactions are not used for training or improvement.

Operational access for support, security, and reliability may still occur when needed, regardless of the training toggle.

Illustrations (Opt-in Sharing)

The only data we actively learn from are user-submitted Illustrations--the Narrative Function examples you choose to share with the community. This is fully opt-in:

• Keep Make All New Illustrations Private enabled to ensure everything you add stays private.
• Disable the setting (or unlock an individual Illustration) when you want to contribute to the shared library.
• Adjust existing Illustrations one by one by toggling the padlock beside each entry.

Training Scope

Training and improvement use is controlled by your settings and is limited to service quality objectives. We do not sell your content or use it for advertising.