bookge ı st
English
English Español

AI image generation safety and suitability

Last modified: 2026-07-12

This page explains how BookGeist’s artificial intelligence image generation works, the controls we apply, and the limitations users should understand. It supplements the BookGeist Privacy Policy and Terms and Conditions.

1. Purpose of the feature

BookGeist can create optional visual references for characters, places, objects, and other reading elements from a description written by the user.

The feature:

  • runs only when the user requests it;
  • is not a social network and does not provide a public gallery of images generated by other users;
  • is not designed to produce pornography, exploitation, graphic or sadistic violence, hate, harassment, dangerous instructions, or illegal content;
  • may occasionally produce non-graphic narrative elements found in fiction, such as fantasy, conflict, creatures, or weapons.

2. Safety controls

A request may pass through several control layers before a result is shown:

  1. BookGeist validation. We check the request format and limits.
  2. BookGeist controls. The request is assessed using moderation and security checks and may be rejected before an image is generated.
  3. AI provider controls. The selected provider and model apply their own policies, classifiers, and safety filters. They may reject either the description or the resulting output.
  4. User review. Generated images are automated proposals. The user decides whether to keep one and can delete it or report an unsuitable result.

A request blocked before execution does not consume generation credits. The exact controls may evolve in response to new risks, models, and legal requirements.

3. Content we do not permit

BookGeist image generation must not be used to create, request, or distribute content that includes or facilitates:

  • child sexual abuse or exploitation;
  • sexualization or nude depictions of minors;
  • non-consensual intimate imagery or sexual violence;
  • pornography or explicit sexual content;
  • graphic, prolonged, or sadistic violence, or content intended to glorify real suffering;
  • terrorism, violent extremism, or credible incitement to violence;
  • hate, harassment, threats, or humiliation directed at people or groups;
  • promotion of or instructions for self-harm;
  • harmful weapons manufacture, procurement, or use;
  • illegal activity or instructions intended to cause harm;
  • deceptive impersonation or unauthorized use of a real person’s likeness;
  • infringement of copyright, privacy, publicity, or other third-party rights;
  • attempts to bypass, abusively test, or evade safety controls.

Artistic, literary, educational, or documentary context may affect how a request is assessed, but it does not remove legal prohibitions or child-safety protections.

4. Providers and models

Depending on the feature and availability, a request may be processed by OpenAI, Google Gemini, OpenRouter, or other providers identified in the Privacy Policy.

Each provider remains responsible for its own models and applies additional terms and controls. Relevant public references include:

Provider selection may change without changing the purpose of the feature. We will update this page when a change materially affects its controls, processing, or risks.

5. Privacy

To generate an image, BookGeist transmits the visual description entered by the user, the necessary style settings, and minimum technical data required to authenticate the request, manage credits, prevent abuse, and deliver the result.

BookGeist does not automatically send the complete library, complete notebook, contacts, backups, or files the user has not selected for this feature.

Details about controllers, purposes, retention, providers, transfers, and rights are available in the BookGeist Privacy Policy.

6. Limitations

Automated generation is not infallible. Even with several layers of controls:

  • a legitimate request may be rejected by mistake;
  • an image may contain unexpected, inaccurate, biased, or unsuitable details;
  • no filter guarantees detection of every possible risk;
  • images must not be treated as facts, professional advice, or authentic depictions of real people or events.

BookGeist may limit, suspend, or disable the feature when abuse is detected, a safety risk exists, or a provider no longer offers sufficient safeguards.

7. Reporting unsuitable content

You can report an image using the controls shown alongside the visual reference. You can also contact support@bookgeist.app.

Reports create a record of the incident so that it can be reviewed, safety rules can be improved, and, where appropriate, results can be removed or abusive use restricted. They are scheduled for automatic deletion after 180 days; temporary failures are retried. We may delete them earlier when no longer needed, and they are deleted when the account is deleted. Do not include unnecessary personal information in a report.

8. Age suitability

BookGeist does not offer violence, weapons, or sexual content as its main purpose. However, because this is a creative tool connected to literary works, some images may occasionally contain fantasy scenes, non-graphic conflict, or weapons in a narrative context.

The rating displayed in an app store is calculated using that platform’s questionnaire and definitions. This page describes the feature and its controls but does not replace the official rating published by each store. See the Age Suitability Policy for BookGeist’s general approach and specific guidance for users aged 13 to 17. You can also consult Apple’s age-rating definitions.

9. Updates

We will review this policy when the feature, providers, controls, or applicable requirements materially change.