Skip to main content
QTPIE
> QTPIE_OS // LEGAL_PROTOCOLS // AI_DISCLOSURES

AI Disclosures

LAST_UPDATED: 2026-04-19VERSION: 2.4READ_TIME: ~2MIN

We believe you deserve to understand how the AI technology powering QtPie actually works, what it can do, and — just as importantly — where it falls short. This document is our honest explanation.

[01]

How Our AI Works (In Plain Language)

When you upload a photograph and select an art style, here is what happens:

  1. Your photograph is transmitted securely to our AI processing provider, fal.ai.
  2. The AI model — FLUX.1 Kontext, developed by Black Forest Labs — analyzes the visual characteristics of your photograph, including your apparent facial features, hair, and overall composition.
  3. The model then generates a new image in your selected art style (anime, chibi, pixel art, cyberpunk, etc.) that interprets your appearance through that artistic lens.
  4. The generated avatar is returned to you and stored in your gallery.
  5. Your original photograph is stored securely in your account so you can reuse it for future avatar generations. You can delete it at any time from your privacy settings (permanent deletion within 24 hours), and it is always deleted when you close your account. fal.ai deletes input images immediately after processing on their end.

The model does not “remember” your face between sessions. Each generation is independent.

[02]

What Your Avatar Is (And Is Not)

Your avatar IS:

  • A stylized, artistic interpretation of the features visible in your uploaded photograph
  • Generated by a neural network trained on large datasets of images
  • An original output created in response to your specific inputs
  • Yours to use as a gaming avatar, profile picture, or personal branding image

Your avatar IS NOT:

  • A photorealistic copy or digital twin of your face
  • A guaranteed accurate representation of your appearance
  • Necessarily unique — other users with similar photos and styles may receive visually similar outputs
  • A biometric record or identity document
  • Guaranteed to be copyrightable under current law (see Section 5)

We do not guarantee that any generated avatar will resemble you, satisfy you aesthetically, or perform identically to examples shown in our marketing materials.

[03]

Known Limitations

3.1 Photo Quality Dependency

Results depend heavily on your input photograph. Low-resolution photos, poor lighting, unusual angles, heavy obstructions (glasses, hats, masks), and motion blur all reduce generation quality. For best results: face the camera directly, ensure even lighting, and use a photo where your face is clearly visible.

3.2 Style Abstraction

Some art styles (especially pixel art, chibi, and highly stylized anime) significantly abstract facial features. In highly stylized outputs, individual features may be difficult to distinguish from other users' avatars in the same style.

3.3 Skin Tone and Ethnicity Representation

We disclose the following limitation transparently: AI image generation models, including the model we use, were trained on large internet datasets that reflect real-world imbalances in image representation. Research has documented that these models can generate outputs that lighten or alter skin tones relative to the input photograph, and produce results that reflect majority-group biases embedded in training data.

We have implemented technical measures in our prompt engineering to mitigate skin tone drift, including skin tone classification and descriptor injection. However, we cannot guarantee equal performance across all skin tones, ethnicities, and facial features. We are actively working to improve representation fairness.

If you notice systematic skin tone or feature alterations in your avatars, please report this to [email protected]. Your feedback directly informs our improvement work.

3.4 Accessories and Attributes

Glasses, jewelry, distinctive hairstyles, and facial hair may or may not be accurately captured depending on the style and the input photo. Reflective or thick-framed eyeglasses may interfere with facial feature recognition.

[04]

How the AI Model Was Trained

The FLUX.1 Kontext model was developed by Black Forest Labs and is accessed through fal.ai. We do not control the model's training data.

QTPIE COM, LLC does not:

  • Use your uploaded photographs to train any AI model without your separate explicit consent
  • Use your generated avatars to train any AI model without your separate explicit consent
  • Provide your photographs to fal.ai or any other party for model training

Your photographs are used for one purpose: generating your avatars. They are stored in your account for as long as you choose to keep them, and are permanently deleted within 24 hours once you delete them from your privacy settings or close your account.

[05]

Intellectual Property and Copyright

Current US law requires human authorship for copyright protection. Courts, including the US Court of Appeals for the D.C. Circuit in Thaler v. Perlmutter (affirmed by Supreme Court denial of certiorari, March 2026), have held that AI systems cannot be copyright authors.

What this means for you:

  • Avatars generated entirely by AI, without significant human creative authorship in the expressive elements, may not be eligible for copyright protection under US law
  • We assign all rights we have to you (see Terms of Service Section 8.3), but this does not guarantee that copyright attaches
  • Other jurisdictions may treat AI-generated content differently; consult a legal advisor if copyright is critical to your use case
[06]

AI-Generated Content Labeling

All avatars generated by our Service are AI-generated images. We plan to embed machine-readable metadata in generated avatar files indicating they were created using AI (Content Credentials / C2PA standard). This feature is planned for a future release. When distributing avatars publicly, you should disclose to your audience that they are AI-generated, as required by applicable platform rules and laws in your jurisdiction.

If you are located in the European Union: the EU AI Act (Article 50) requires that AI-generated content that could be mistaken as depicting real people must be labeled as artificially generated. While stylized gaming avatars are generally understood to be artistic representations, we encourage transparent disclosure of their AI-generated nature.

[07]

Do Not Use for Identity Verification

Avatars generated by QtPie should not be used for:

  • Identity verification or authentication
  • Legal identification documents
  • Biometric matching or security screening
  • Any context where an accurate photographic likeness is required

Avatars are artistic interpretations, not identity documents.

[08]

Our Commitment to Improvement

We are committed to:

  • Continuously improving fairness and representation across skin tones and ethnicities
  • Monitoring for and mitigating harmful outputs
  • Staying current with regulatory requirements including the EU AI Act
  • Being transparent with you about how our technology works and where it falls short

If you experience issues or have feedback about AI performance, fairness, or output quality, please contact us at [email protected].

[09]

Contact

QTPIE COM, LLC

Email: [email protected] | [email protected]

Website: QtPie.com/legal/ai-disclosures

END_OF_DOCUMENT // AI_DISCLOSURES // v2.4
↑ BACK_TO_TOP