AI Transparency

1. AI Technology We Use

1.1 AI Provider

We use Anthropic's Claude API, a general-purpose AI system, to generate content suggestions. Claude is designed with safety and helpfulness as core principles. For more information about the model, visit anthropic.com.

1.2 Risk Classification

Under the EU AI Act's risk-based framework, FeedMansion's AI features are classified as limited risk AI systems. This classification applies because:

• The AI generates content suggestions, not autonomous decisions
• All AI output requires human review before publication
• Users maintain full editorial control
• The system does not make decisions affecting fundamental rights

As a limited-risk system, we are subject to transparency obligations under Article 50 of the AI Act, which this document fulfills.

2. AI-Powered Features

FeedMansion uses AI in the following specific features:

3. Human Control and Oversight

3.1 The Review Process

1. Generation: AI creates a draft based on your content and Ghost personality
2. Labeling: All AI-generated drafts are clearly marked with an "AI generated" badge
3. Review: You see the draft in your Review queue
4. Edit: You can modify any part of the draft before approval
5. Decision: You approve, reject, or edit each draft individually
6. Publication: Only approved content is scheduled or published

3.2 Your Editorial Responsibility

Because you review and approve all content before publication, you hold editorial responsibility for published posts. Under Article 50(4) of the EU AI Act, content that undergoes human review and editorial control does not require AI disclosure to the public when published.

However, within the FeedMansion app, all AI-generated content remains clearly labeled so you always know its origin.

4. Data Processing for AI Features

4.1 What Data is Sent to the AI

When you use AI features, the following data is sent to Anthropic's API:

• For draft generation: RSS article title, link, and description; your Ghost's name and personality description; your preferred posting language
• For alt text: The image file (temporarily, for processing only)
• For calendar posts: Event title, date, location, and description

4.2 What Data is NOT Sent

• Your personal account information (email, name)
• Your social media account credentials
• Your payment information
• Content from other users

4.3 Data Retention by AI Provider

According to Anthropic's data usage policy, API inputs and outputs are not used to train their models. Data is processed for the request and not retained beyond what is necessary for service delivery and abuse prevention.

For full details, see Anthropic's Privacy Policy.

5. Your Rights and Controls

6. AI Limitations and Risks

AI-generated content has inherent limitations. We are transparent about these:

• Accuracy: AI may occasionally produce factually incorrect information. Always verify important claims.
• Tone: AI interpretation of your Ghost's personality may not always match your expectations.
• Context: AI lacks real-world context and may miss nuances that are obvious to humans.
• Originality: AI-generated text is based on patterns in training data and may occasionally produce similar content to other sources.
• Platform policies: AI cannot guarantee compliance with social media platform rules. You must review for compliance.

This is why human review is mandatory. The AI assists; you decide.

7. Legal Framework

7.1 EU AI Act Compliance

This disclosure is provided pursuant to Article 50 of Regulation (EU) 2024/1689 (the AI Act). As a provider and deployer of AI systems that interact with users and generate content, we are committed to transparency about:

• The fact that users are interacting with AI-generated content (Article 50(1))
• Clear labeling of AI-generated outputs within our application (Article 50(2))
• The human review process that applies before publication (Article 50(4))

7.2 GDPR Compliance

AI-related data processing is conducted in accordance with the General Data Protection Regulation (GDPR). Legal bases include:

• Art. 6(1)(b) GDPR: Processing necessary for contract performance (providing the AI features you use)
• Art. 6(1)(f) GDPR: Legitimate interests (service improvement, fraud prevention)

For full details on data processing, see our Privacy Policy.

8. Updates to This Disclosure

We will update this AI Transparency disclosure when:

• We change AI providers or models
• We add new AI-powered features
• Regulatory requirements change
• We make significant changes to how AI data is processed

Material changes will be communicated to registered users via email.