AI Disclosure & Transparency

v1.0 · Last updated: 2026-03-18

This page explains how Writegarden uses artificial intelligence, in compliance with the EU AI Act (Regulation (EU) 2024/1689) — specifically the transparency obligations under Article 50 for deployers of AI systems that generate synthetic content. The EU AI Act Article 50 transparency obligations become enforceable on 2 August 2026.

1. Our AI Classification

Under the EU AI Act, Writegarden is classified as follows:

  • Risk level: Limited risk (not high-risk under Annex III)
  • Role: Deployer of general-purpose AI systems — we use third-party AI models via API to generate content
  • Article 50 applicability: Yes — we generate synthetic text and images intended for public publication

2. AI Systems We Use

The following AI models and systems are used to power our Service:

Article Text Generation

Generates SEO-optimized blog articles based on your business context, target keywords, and competitor analysis.

  • Models: GPT-4, Claude
  • Accessed via: OpenRouter API
  • Input data: Business context, keywords, competitors, tone preferences
  • Output: HTML article content, meta descriptions, SEO titles

Domain Research

Researches your domain and business to build an accurate company profile for content generation.

  • Model: Perplexity
  • Accessed via: OpenRouter API
  • Input data: Website URL, domain name
  • Output: Business description, industry context

Image Generation

Creates original illustrations for articles based on the article's topic and your style preferences.

  • Model: Gemini 2.5 Flash
  • Accessed via: OpenRouter API
  • Input data: Article context, style description
  • Output: PNG images

Keyword Intelligence

Suggests negative keywords and filters irrelevant keywords from your research results.

  • Model: Custom AI via OpenRouter
  • Input data: Keyword lists, business context
  • Output: Filtered keyword recommendations

3. AI-Generated Content Notice

All content produced by Writegarden — including articles, meta descriptions, titles, and images — is generated by artificial intelligence. This content:

  • Is not written, reviewed, or verified by humans before being presented to you in the platform
  • May contain factual errors, inaccuracies, or outdated information
  • May resemble existing third-party content due to the nature of AI training data
  • Should be treated as a first draft that requires your review and editing

You are responsible for reviewing, fact-checking, and editing all AI-generated content before publication. We strongly recommend that a human with subject-matter expertise reviews every piece of content before it is published.

4. How We Mark AI-Generated Content

In compliance with Article 50 of the EU AI Act, we take the following measures to mark AI-generated content:

4.1 Within the Platform

  • All generated articles are clearly labeled as "AI-Generated" in the article editor
  • A persistent notice is displayed when viewing or editing AI-generated content
  • The publish flow includes a confirmation step acknowledging the AI-generated nature of the content

4.2 Machine-Readable Metadata

  • Generated articles include machine-readable metadata indicating AI origin (e.g., <meta name="ai-generated" content="true">)
  • AI-generated images include metadata tags indicating synthetic origin where technically feasible

4.3 Published Content

When publishing content to third-party CMS platforms (Webflow, Shopify, WordPress, Wix, Framer), we offer the option to include an AI disclosure footer in the published article. The standard disclosure reads: "This article was created with AI assistance and reviewed by [your company name]."

5. Your Obligations as a Content Publisher

When you publish AI-generated content from our platform, you become a deployer of AI-generated content under the EU AI Act. This means you have certain obligations:

  • Inform your audience: Under Article 50(4) of the EU AI Act, deployers of AI systems that generate synthetic text published to inform the public must disclose that the content has been artificially generated or manipulated. You should inform readers that published articles were created with AI assistance.
  • Review before publishing: You must review all AI-generated content for accuracy, completeness, and appropriateness before making it public.
  • Label AI images: Under Article 50(2), AI-generated images must be clearly labeled or watermarked as artificially generated when made available to the public.
  • Maintain machine-readable markers: Where technically feasible, preserve the machine-readable metadata indicating AI-generated content.

6. Data Sent to AI Providers

When generating content, we send the following types of data to AI providers through server-side requests (never directly from your browser):

  • Business context (company name, industry, description, tone, target audience)
  • Target keywords and SEO parameters
  • Competitor information
  • Style preferences for image generation

We do not send to AI providers your email address or account credentials; OAuth tokens or integration credentials; payment or billing information; or other users' data.

7. Known Limitations of AI Systems

We are transparent about the limitations of the AI systems we use:

  • Hallucination: AI models may generate plausible-sounding but factually incorrect information. Always verify factual claims.
  • Bias: AI models may reflect biases present in their training data. Review content for unintended bias.
  • Recency: AI models have knowledge cutoff dates and may not reflect the most current information or events.
  • IP overlap: AI-generated content may inadvertently resemble existing copyrighted works. Check for uniqueness before publishing.
  • Context sensitivity: AI may not fully understand industry-specific nuances, regulations, or cultural context relevant to your business.

8. Questions About Our AI Usage

If you have questions about how we use AI, the data we process, or your obligations under the EU AI Act, contact privacy@writegarden.com.