AI Customer Data Policy

1. Purpose of the Policy

This policy describes how TAGMOOD S.R.L. uses artificial intelligence technologies within Webmaster360 and how Customers’ data and content are processed during the use of the relevant features.

Artificial intelligence is used exclusively as a support tool for the work of editors, journalists, authors and other authorized users of the platform.

Webmaster360 does not use artificial intelligence to:

• profile readers;
• determine which individuals may access the service;
• make automated decisions with legal effects on individuals;
• evaluate employees, collaborators or candidates;
• assign reliability scores to individuals;
• independently determine editorial purposes;
• create its own advertising profiles;
• independently decide which content to publish.

A functional description of editorial automations is available on the AI and Editorial Automation page.

2. Role of Artificial Intelligence

Webmaster360’s AI features are designed to speed up repetitive editorial tasks and help the editorial team prepare, adapt and organize content.

Artificial intelligence may be used, depending on the features activated by the Customer, to:

• rewrite content imported from feeds or authorized sources;
• transform notes, press releases or raw text into an editorial draft;
• generate title, subtitle and article body;
• adapt tone, structure and editorial style;
• process content starting from one or more source links;
• duplicate and adapt an article for another site within the same network;
• prepare recurring content, such as weather bulletins, anniversaries or rankings;
• create texts intended for publication or sharing on configured social channels;
• support further editorial automations requested by the Customer.

Not all automations necessarily use artificial intelligence. Some activities may consist of imports, scheduling, technical processing or rule-based publications.

3. Data and Content That May Be Processed

When an AI feature is used, Webmaster360 may process and transmit to the AI provider the data strictly necessary to execute the request.

This data may include:

• prompts and instructions entered by the authorized user;
• notes, press releases or source texts;
• selected articles, drafts or excerpts;
• content acquired from feeds, APIs or configured sources;
• text extracted from source links indicated by the user;
• title, subtitle, body and editorial metadata;
• indications relating to tone, structure, style or format;
• contextual information added by Webmaster360 to obtain the requested result;
• outputs generated by the model;
• technical metadata necessary for the execution and security of the request.

Webmaster360 does not intentionally send the AI provider data beyond what is necessary for the requested function.

The Customer and authorized users determine which texts, links, sources and data are submitted for processing.

4. AI Provider Used

The provider currently used is:

OpenAI Ireland Limited
Service: OpenAI API

OpenAI is listed in the sub-processor list.

Webmaster360 uses the provider’s API and not users’ personal ChatGPT accounts.

TAGMOOD may replace or add a provider in accordance with the methods provided for by the Data Processing Agreement and the sub-processor update procedure.

5. Use of Data for Training

TAGMOOD does not use Customer Data to train its own or third-party artificial intelligence models.

Under the terms applicable to the OpenAI API platform, inputs and outputs sent via API are not used to train or improve models, unless the customer explicitly opts in to data sharing.

TAGMOOD does not voluntarily opt in to provider programs that provide for the use of Customer Data for training, unless otherwise instructed in writing by the Customer.

6. Retention at the Provider

In the ordinary API configuration, OpenAI may generate abuse monitoring logs containing prompts, responses and related metadata.

According to the provider’s documentation, such logs may be retained for up to 30 days, unless:

• longer retention is required by law;
• it is reasonably necessary to protect the services or third parties;
• different retention configurations approved by the provider have been activated.

Webmaster360 does not currently declare the activation of Zero Data Retention for all AI requests.

Any data maintained by Webmaster360 in the CMS, such as drafts, articles or outputs saved by the user, follows the platform’s retention rules and the Customer’s instructions, not the technical term applied by the AI provider to its own logs.

7. Editorial Control and Publication Methods

Each workflow can be configured by the Customer based on the desired level of control.

The Customer may choose, according to the available features, to:

• import content without rewriting;
• rewrite it with artificial intelligence;
• save it as a draft;
• submit it for editorial review;
• schedule its publication;
• publish it automatically;
• share it automatically on configured social channels.

Human review before publication therefore depends on the configuration chosen by the Customer.

Webmaster360 provides control tools, but:

• the selection of sources remains the editorial team’s responsibility;
• fact-checking remains an editorial responsibility;
• the decision to publish remains with the Customer;
• the Customer determines which workflows may be automated;
• the Customer is responsible for the automatic publication configuration.

For sensitive or high-risk content, TAGMOOD recommends configuring the saving as a draft and an editorial review before publication.

8. Accuracy and Limitations of Outputs

Artificial intelligence systems may:

• produce inaccurate information;
• misinterpret a source;
• omit relevant information;
• generate formulations not supported by the sources;
• introduce errors, distortions or repetitions;
• create content not suitable for the editorial context.

Outputs must not be considered automatically accurate, complete or ready for publication.

Before publication, the editorial team should verify at least:

• accuracy of facts;
• consistency with sources;
• attribution of statements and quotes;
• names, dates, places and numerical data;
• presence of personal information;
• content and image rights;
• tone and editorial appropriateness;
• any risks to individuals or groups.

Artificial intelligence should be used to speed up work, not to reduce quality standards.

9. Sources, Permissions and Content Ownership

The Customer is responsible for using:

• authorized feeds;
• agencies and editorial partners that are lawfully accessible;
• websites and sources that may be lawfully monitored or acquired;
• texts, documents, images and other materials for which it holds the necessary rights or legal bases.

The technical possibility of acquiring, analyzing or rewriting content does not constitute authorization for its use.

The use of multiple sources or AI rewriting does not eliminate obligations relating to:

• copyright;
• attribution;
• licenses;
• confidentiality;
• data protection;
• right to report;
• accuracy of information.

10. Sensitive Content and YMYL

The Customer must not automate, without adequate editorial control, content that may significantly affect the health, safety, financial situation, reputation or rights of individuals.

Particular caution is required for content relating to:

• health and medicine;
• emergencies and security;
• judicial proceedings and criminal offences;
• minors;
• sensitive personal data;
• politics, elections and civic matters;
• finance, investments and taxation;
• information that may cause significant harm if inaccurate.

In such cases, the Customer should:

• use reliable and up-to-date sources;
• verify the information with competent persons;
• maintain human review;
• avoid unsupervised automatic publication;
• limit personal data sent to the AI;
• document the sources and verifications carried out.

11. Special Categories of Data, Criminal Offence Data and Minors

Users must not enter special categories of personal data, data relating to criminal offences or data relating to minors into the AI features, unless:

• the processing is strictly necessary;
• a valid legal basis exists;
• the use is consistent with the editorial purpose;
• adequate protective measures have been adopted;
• the Customer has assessed the risks of the processing.

Where possible, personal data should be removed, anonymized or minimized before being sent to the AI provider.

Secrets, credentials, API keys, passwords or confidential information not necessary for the requested function must not be entered.

12. Automatic Publication and Social Sharing

Webmaster360 may allow the Customer to link the generation or import of content to automatic publication and, if configured, to sharing on social media.

These functions are controlled by the Customer’s configurations.

Before activating them, the Customer should assess:

• reliability of the source;
• frequency and predictability of the content;
• risk of errors;
• possible impact on individuals;
• need for review;
• presence of personal data;
• risk of massive or duplicative publication;
• procedures for correction, suspension and removal.

Low-risk, predictable automations based on structured data may require a different level of control compared to sensitive news, current affairs or content based on unverified sources.

13. Disabling and Configuration

AI features can be disabled.

The Customer may define:

• which users may use them;
• which sources or workflows may be processed;
• whether the content should remain as a draft;
• whether a review is required;
• whether a workflow may publish automatically;
• whether publication should trigger social sharing.

For requests relating to the configuration of AI features, please contact:

webmaster360@tagmood.it

14. Privacy Roles

Where TAGMOOD processes content and data through AI features to provide the service requested by the Customer, TAGMOOD generally acts as a data processor on behalf of the Customer.

The Customer acts as controller and determines:

• purposes of the processing;
• sources and content used;
• personal data submitted for processing;
• review methods;
• publication methods;
• retention periods in the CMS.

OpenAI acts as a sub-processor as described in the Data Processing Agreement and the sub-processor list.

Further information is available in:

• Data Processing Agreement;
• Privacy & Data Protection;
• Sub-processors.

15. Security

Communications between Webmaster360 and the AI provider are protected through encrypted protocols.

Access to AI features is reserved for users authorized by the Customer.

TAGMOOD applies technical and organizational measures aimed at protecting:

• provider credentials;
• API requests;
• data stored in the CMS;
• logs and technical information;
• Customer data segregation.

Further information is available in the:

• Security Overview;
• Description of Technical and Organizational Measures.

16. Editorial Responsibility

Editorial responsibility always remains with the Customer.

Webmaster360 provides tools to:

• acquire content;
• organize it;
• process it;
• generate versions of it;
• save it as a draft;
• publish it;
• distribute it.

The Customer decides how to use such tools and remains responsible for:

• the accuracy of the content;
• the choice of sources;
• compliance with rights and permissions;
• verification before publication;
• the configuration of automations;
• any corrections and removals that may be necessary;
• the training of its editorial team.

17. Changes to the Policy

TAGMOOD may update this policy to reflect changes to:

• AI features;
• the providers used;
• processing configurations;
• Applicable Law;
• security and retention practices.

Each version shall bear the effective date and last update date.

Changes concerning the addition or replacement of a sub-processor shall follow the notice and methods provided for by the Data Processing Agreement.

18. Contacts

For information on this policy or on AI features:

• Privacy: privacy@tagmood.it
• Support and configuration: webmaster360@tagmood.it
• PEC: tagmood@pec.it

Related Documents

• Data Processing Agreement
• Privacy & Data Protection
• Sub-processors