AI Transparency & Policy
Last updated: April 2026
Publica.la integrates AI capabilities into the reading experience to help readers comprehend, interact with, and get more value from digital content. This page explains what AI features exist, how content is processed, what we do and do not do with publisher data, and what controls are available.
How We Use AI
All AI features are reader-facing tools, activated on-demand by the end user during a reading session. No AI processing occurs in the background or without explicit user action.
We make every effort to keep this list current, but it may not reflect the very latest additions at all times. Regardless, the commitments on this page regarding how content is processed, stored, and handled will not change without prior notice to publishers.
Standard Reader Tools
Standard digital reading tools present in most modern reading platforms. Always available to readers.
| Feature | Description |
|---|---|
| Dictionary | Provides definitions and part-of-speech information for selected words |
| Translate | Translates selected text to the reader's preferred language |
| Text-to-Speech | Converts text to spoken audio using synthetic voice |
AI Comprehension Tools
AI-powered tools that work on selected text (and in some cases, page ranges) to help readers understand and interact with content. Enabled by default; publishers can request to disable them.
| Feature | Description |
|---|---|
| Explain | Reader selects a passage and receives a simplified explanation |
| Expand | Provides additional context and detail on a selected passage |
| Summarize | Generates a summary of selected text or page ranges |
Vito AI Assistant
Vito is a conversational AI assistant embedded in the reader. Available on request; publishers can activate it for their catalog.
Vito generates summaries of whole or multiple pages, creates assessment questions from content, and answers reader questions based on the publication text.
How Your Content Is Processed
- On-demand only: AI processing occurs only when a reader explicitly triggers a feature (e.g., selects text and taps "Explain"). There is no background processing or batch analysis.
- Minimal data sent: Only minimal text content, metadata (language, book title), and reader questions (for Q&A features) are sent to the AI provider. No reader identifiers or personally identifiable information are transmitted.
- Encrypted in transit: All data is encrypted in transit over TLS 1.2+ connections.
Our Commitments
Publica.la is committed to the following practices regarding publisher content:
- Publisher content is not used to train, fine-tune, or improve any AI models.
- AI outputs (explanations, translations, summaries) function as reading aids for the individual reader and are not stored or redistributed.
- Publisher content is not shared with third parties beyond the API calls described above.
- AI is used solely to enhance the individual reader's experience with the content they already have authorized access to.
Publisher Controls
- Standard Reader Tools (Dictionary, Translate, Text-to-Speech) are always available as part of the reading experience.
- AI Comprehension Tools (Explain, Expand, Summarize) are enabled by default and can be disabled per publisher.
- Vito AI Assistant is available on request and can be activated per publisher.
To adjust AI settings for your catalog, contact your account manager or reach out to support@publica.la.
AI Providers & Compliance
We work with enterprise-grade AI providers that maintain strict data handling policies:
| Provider | Compliance |
|---|---|
| OpenAI | SOC 2 Type II, GDPR compliant, zero data retention on API inputs |
| Anthropic | SOC 2 Type II, GDPR compliant, zero data retention on API inputs |
| Amazon Web Services (Polly) | SOC 2, ISO 27001, HIPAA eligible, GDPR compliant |
All providers operate under data processing agreements (DPAs) that prohibit them from using API inputs for model training or any purpose other than fulfilling the specific request.
What Is Not AI
For clarity, the following platform features do not currently use AI, machine learning, or any form of computational intelligence:
- Search: Uses traditional full-text database search on indexed metadata fields.
- Recommendations: Uses rule-based matching on BISAC (Book Industry Standards and Communications) category taxonomy. No machine learning is involved.
- Content intake and distribution: ONIX file processing and PDF optimization are purely mechanical operations with no AI component.
Frequently Asked Questions
Does Publica.la use publisher content to train AI models?
No. Publisher content is not used for training, fine-tuning, or improving any AI model. Our AI providers contractually commit to not using API inputs for model training.
Is reader data shared with AI providers?
No reader identifiers or personally identifiable information are sent to AI providers. Only minimal text content, metadata (such as language preference), and reader questions (for Q&A features) are transmitted.
Which AI features can publishers control?
AI Comprehension Tools (Explain, Expand, Summarize) can be disabled per publisher. Vito AI Assistant is only active when explicitly enabled. Standard Reader Tools (Dictionary, Translate, Text-to-Speech) are part of the core reading experience. Contact your account manager or support@publica.la to adjust settings.
Does this comply with GDPR?
Yes. All AI providers maintain GDPR-compliant data processing agreements. No personal data is sent to AI providers. For details on our broader privacy practices, see our Privacy Policy.
Are AI-generated outputs stored or redistributed?
AI outputs (explanations, translations, summaries, etc.) are delivered to the reader and are not redistributed by Publica.la.
Where can I find more technical details?
For a detailed technical overview including data flow diagrams and security architecture, see our AI Transparency & Data Handling documentation.
Contact
If you have questions about our AI practices or need additional documentation for your agreements, please contact us at support@publica.la.