AI CONTENT DISCLOSURE

Last Updated: March 6, 2026

We believe in being upfront about how our platform works. This page explains how images on this platform are created, what role human judgment and editing play in the process, what AI-generated means for copyright, and what you should know before using images in your projects.

This page is written in plain language. The full legal terms governing your use of images are in our Terms and Conditions and Image License Agreement.

1. How Images Are Created on This Platform

Images on this platform are not photographs. They are created using artificial intelligence tools and then reviewed and edited by a human before being published.

Our image creation process involves three stages:

Stage 1: AI Generation

Images are initially generated using one or more AI platforms.

Each platform uses its own AI models trained on large image datasets. Links to each platform’s terms and data policies are provided in Section 2.

Stage 2: Human Review

Every AI-generated image is reviewed by our Creative Director before it is published to the library. Images that do not meet our quality standards, that contain inappropriate content, or that raise concerns about resemblance to real individuals or existing works are not published and are discarded.

Stage 3: Human Editing

Many images undergo additional editing in Adobe Photoshop before being published. This may include color correction, composition adjustments, removal of artifacts or errors introduced by the AI generation process, and other refinements. This human creative input is applied to improve quality and ensure images are fit for professional use.

* The combination of AI generation, human review, and human editing means our images are not purely machine-made. Human creative judgment is applied at every stage of the process.

What This Means for Copyright

The involvement of human creative editing at the Photoshop stage may support intellectual property rights in the final edited versions of images, distinct from the raw AI-generated output. This is an evolving area of law. Our Image License Agreement governs your usage rights regardless of copyright status. See Section 3 for more detail.

2. AI Platforms & Policies

We use multiple AI platforms to generate images. Each platform has its own terms of service, content policies, and training data practices. The table below lists each platform we use and links to their relevant policies. We encourage you to review these if you have questions about training data or model behavior.

Platform	Primary Use	Policy URL
Midjourney	Image generation	midjourney.com/tos
Freepik	Image generation	freepik.com/terms-of-use
ChatGPT / DALL-E	Image generation	openai.com/policies
ElevenLabs	Image generation	elevenlabs.io/terms
Adobe Photoshop	Post-production editing	adobe.com/legal

Note: Photoshop is a human editing tool, not an AI generation platform. It is listed above for transparency about our full production process. Adobe Photoshop does include some AI-assisted features (such as Generative Fill) which may be used during editing.

ElevenLabs is used on this platform for AI image generation via their text-to-image tool.

3. What AI-Generated Means for Copyright

3.1 Copyright Status of AI Images

Under current guidance from the U.S. Copyright Office, images created purely by artificial intelligence without sufficient human creative authorship are generally not eligible for copyright protection. This is a developing area of law that varies by country and is actively being reviewed by courts and regulators worldwide.

Because our images involve human creative editing in Photoshop, there may be grounds for intellectual property protection in the final edited versions. However, we do not make a blanket copyright claim over all images on the platform, as the level of human input varies by image.

3.2 Your Rights Are Contractual

Regardless of copyright status, your right to use images from this platform is real and legally binding. It is governed by a contractual license, our Image License Agreement, which you agree to when you subscribe. That agreement defines what you can and cannot do with images and is enforceable independently of copyright law.

3.3 Laws Are Changing

AI and copyright law is one of the fastest-moving areas of intellectual property globally. Courts and legislators in the US, EU, and UK are actively working through these questions. We monitor these developments and will update our policies as the legal landscape evolves.

4. Realistic Faces & People in AI Images

AI-generated images may depict realistic human faces and figures. These are entirely synthetic as no real person was photographed. However, AI can occasionally produce likenesses that resemble real individuals by coincidence.

Here is what you should know:

Faces and figures in our images are algorithmically generated and are not intended to depict real, identifiable individuals.
Our human review process screens for images that appear to closely resemble known public figures or individuals, and such images are not published.
Despite this review, we cannot guarantee that every generated face bears no resemblance to any living or deceased person.
If you use images containing realistic faces in advertising, endorsement, or political contexts, you assume responsibility for any claims related to likeness rights or right of publicity.
We strongly recommend against using AI-generated faces in any context where a specific individual’s endorsement or identity could be implied.

5. Training Data & Third-Party Content

The AI models we use are trained on large datasets of existing images. Each platform’s approach to training data is governed by their own policies, linked in Section 2. In general:

Midjourney trains on a large proprietary dataset. Details are available in their terms of service.
Freepik’s Nano Banana and Seedream models are trained on datasets governed by Freepik’s content policies.
OpenAI’s DALL-E 3, used via ChatGPT, is trained on data subject to OpenAI’s usage policies.
ElevenLabs’ text-to-image models are trained on datasets governed by ElevenLabs’ content and usage policies.

What this means for you:

It is theoretically possible for AI models to generate outputs that bear similarity to existing copyrighted works, though this is not intended and is generally rare.
Our human review process is an additional safeguard as reviewers are instructed to flag and discard any image that appears to replicate a recognizable existing work.
If you believe an image in our library reproduces a specific copyrighted work, please contact us at [email protected] and we will investigate and remove it if appropriate.
Our Image License Agreement includes indemnification provisions covering third-party similarity claims for unmodified images used within your license terms.

6. Your Disclosure Obligations as a Publisher

* You may be legally required to disclose that content you publish was created with AI. This obligation rests with you as the publisher — not with us.

6.1 Why This Matters

A growing number of laws, platform policies, and industry standards require that AI-generated content be clearly labeled. Requirements are expanding rapidly, particularly under the EU AI Act, in various US states, and across major social media and advertising platforms.

6.2 Where Disclosure May Be Required

Social media platforms — Meta, TikTok, YouTube, and X/Twitter have or are developing AI content labeling requirements
Digital advertising networks — Google, Meta Ads, and others may require disclosure of AI content in ad creative
Political advertising in many jurisdictions
Consumer-facing marketing in regulated industries such as financial services and healthcare
News and editorial publications

6.3 What You Should Do

Check the AI content disclosure policies of any platform where you intend to publish;
Check applicable laws in the territories where your content will be seen;
When in doubt, disclose. A simple label such as “Created with AI” is increasingly standard practice and builds audience trust.
If you are on the Agency Plan and share images with clients, make sure your clients are also aware the images are AI-generated and understand their own disclosure obligations.

6.4 Our Position

We are not responsible for your compliance with AI disclosure requirements. We provide this information to help you make informed decisions, but the disclosure obligation rests with the publisher.

7. Our Quality & Safety Standards

Every image is reviewed by a human before it is published. Images that do not meet our standards are discarded.

Our review and editorial process includes:

Human review of every AI-generated image by a member of our team prior to publication;
Human editing in Adobe Photoshop to correct quality issues, remove artifacts, and refine composition and color;
Screening for content that depicts explicit sexual content, graphic violence, or any material that could constitute child sexual abuse material (CSAM), such images are immediately discarded and never published;
Screening for images that appear to closely depict identifiable real individuals, particularly public figures, in misleading or harmful contexts;
Content filters built into the AI platforms we use, which are designed to reduce the generation of harmful or infringing outputs.

Despite these measures, AI generation is an imperfect process and human review has limitations. If you encounter an image in our library that you believe is inappropriate, harmful, or potentially infringing, please report it to us at [email protected] and we will review it promptly.

8. Questions & Contact

If you have questions about how our images are created, your rights as a subscriber, or anything else covered on this page, we are happy to help.

FRILLS

Email: [email protected]

Website: www.getfrills.com

This AI Content Disclosure should be read alongside our Terms and Conditions, Image License Agreement, and Privacy Policy. Together, these documents govern your use of the Frills platform.