AI System Disclosure for Guilted

1. Use of AI in the Service

Guilted uses AI technologies to:

• analyze user-provided descriptions of the gift recipient
• interpret preferences, traits, and personality indicators
• generate creative gift ideas
• assess uniqueness, practicality, and relevance
• optionally propose products through affiliate programs

AI is used only to assist users with recommendations and is not used to make decisions with legal or significant personal effect.

2. AI Providers

We use AI APIs from third-party providers, including:

• OpenAI (GPT-series models)
• Anthropic (Claude models)
• Potential additional LLM providers as the service grows

These providers process text input solely to return generated content.

We do not provide these AI providers with:

• payment data
• behavioral analytics
• tracking identifiers

3. Categories of Data Sent to AI Models

We send only the text provided in the form, such as:

• information about the gift recipient
• optional notes or preferences
• context needed to generate ideas

We do not send:

• full user accounts
• email addresses
• analytics data
• cookies
• metadata not required for service use

4. How AI Generates Gift Ideas

The AI uses:

• natural language understanding
• reasoning models
• creativity prompts
• inference rules
• heuristics and ranking methods

The output is:

• non-binding
• potentially imperfect
• potentially inappropriate in rare cases
• not guaranteed to be correct or reliable

AI suggestions are entertainment and inspiration, not professional advice.

5. Risks & Limitations

AI systems may:

• generate incorrect or unsuitable suggestions
• misunderstand nuanced descriptions
• produce biased outputs
• recommend unavailable or mismatched products
• reflect limitations of their training data

We actively monitor outputs, but complete error elimination is not possible.

6. Human Oversight

Although gift ideas are generated by AI, Guilted ensures responsible use by:

• designing safe prompts
• filtering harmful or explicit content
• monitoring usage patterns
• updating content moderation policies

There is no automated decision-making with legal consequences.

7. User Responsibilities

Users must:

• avoid entering sensitive personal data
• avoid inputting confidential, private, or secret information
• understand that results are suggestions only
• verify any product details directly with third-party shops

8. User Rights

Users may request:

• access to their data
• deletion of account and stored inputs
• correction of information
• restrictions on processing

Requests: mario@36chambers.studio

9. AI Infrastructure & Security

We use:

• encrypted communication
• secure hosting
• role-based internal access
• data minimization principles
• pseudonymization where possible

AI data is not used for training unless explicitly stated by the provider and permitted by user consent.