AI Governance & Responsible AI Policy

FARENIO Ecosystem — Terms, Policies & Agreements

FARENIO ECOSYSTEM

AI GOVERNANCE & RESPONSIBLE AI POLICY

Last updated: 19 September 2026

This Policy is issued by GQ VENTURES SAS, trading as FARENIO, under French and European Union law, and forms Level 2 of the FARENIO Governance Hierarchy set out in Article 1 below.

ARTICLE 1 – GOVERNANCE HIERARCHY

This Policy forms part of the FARENIO AI Governance & Compliance Framework, which establishes the governance architecture applicable to trust, safety, artificial intelligence, information security, platform verification, financial-crime controls and tax-reporting controls throughout the FARENIO Marketplace and Ecosystem.

FARENIO operates the following governance hierarchy:

Level 1 — Trust & Safety

Protection of Clients, Hosts, Service Providers, Cleaners, Vendors, Partners, properties and the integrity of the Marketplace.

Level 2 — Artificial Intelligence Governance

Governance of AI systems, automated decision-making, algorithms and AI-assisted operations.

Level 3 — Security / ISMS

Protection of information, systems, infrastructure, applications, accounts, APIs and business operations.

Level 4 — Identity, KYC/KYB, AML/CFT Interface & DAC7

Verification of marketplace participants, trader traceability, tax-reporting due diligence, payment-provider compliance interfaces and financial-crime risk controls. This layer is addressed in FARENIO's dedicated AML, KYC, CTF & DAC7 Compliance Policy and is not duplicated in this document.

These layers form a single control architecture but remain legally distinct. Compliance with one layer shall not be treated as compliance with another.

This Policy sets out Level 2 — Artificial Intelligence Governance. Trust & Safety controls (Level 1) are addressed in FARENIO's Marketplace Trust, Safety & Integrity Policy, information-security controls (Level 3) are addressed in FARENIO's Information Security Management System Policy, and identity verification, KYC/KYB, AML/CFT and DAC7 tax-reporting controls (Level 4) are addressed separately in FARENIO's AML, KYC, CTF & DAC7 Compliance Policy.

ARTICLE 2 – PURPOSE

This Policy governs the design, acquisition, development, integration, testing, deployment, operation, monitoring, modification and retirement of Artificial Intelligence systems within the FARENIO Ecosystem.

Its objectives are to:

  • Ensure Compliance With French and European Law;
  • Implement Applicable Requirements Of Regulation (EU) 2024/1689;
  • Promote Lawful And Trustworthy AI;
  • Protect Fundamental Rights;
  • Protect Personal Data;
  • Establish Accountability;
  • Maintain Meaningful Human Oversight;
  • Prevent Prohibited AI practices;
  • Identify And Mitigate AI risks;
  • Ensure Transparency;
  • Maintain Security, Robustness And Reliability; And
  • Support Responsible Innovation.

ARTICLE 3 – LEGAL FRAMEWORK

AI activities shall be assessed against applicable requirements including:

  • Regulation (EU) 2024/1689 — Artificial Intelligence Act;
  • Regulation (EU) 2016/679 — GDPR;
  • Loi Informatique et Libertés;
  • Code de la consommation;
  • Applicable Labour And Employment Rules;
  • Digital Services Act Where Applicable;
  • Cybersecurity requirements;
  • Intellectual-property law;
  • Database rights; and
  • Requirements and decisions of competent authorities.

Multiple applicable regimes shall apply cumulatively.

ARTICLE 4 – AI SCOPE

This Policy applies to AI used within:

  • Marketplace;
  • Client Portal;
  • Host Portal;
  • Service Provider/Cleaner Portal;
  • Vendor Portal;
  • Partner interfaces;
  • Admin Portal;
  • SaaS Platform;
  • Property Operations Platform;
  • PMS integrations;
  • APIs;
  • Customer support; and
  • Internal operations.

Covered functionality includes:

  • Pricing Assistance;
  • Scheduling Optimisation;
  • Dispatch;
  • Matching;
  • Operational Intelligence;
  • Customer-Support Assistants;
  • Workflow Automation;
  • Predictive Analytics;
  • Recommendations;
  • Fraud Detection;
  • Quality Analytics;
  • Document Processing;
  • Demand Forecasting;
  • Property Optimisation; And
  • Generative AI.

ARTICLE 5 – AI ACTOR CLASSIFICATION

For each AI system, FARENIO shall determine whether it acts as:

  • Provider;
  • Deployer;
  • Importer;
  • Distributor; or
  • another regulated operator.

Use of a third-party model or API does not eliminate obligations independently applicable to FARENIO.

ARTICLE 6 – RESPONSIBLE AI PRINCIPLES

FARENIO AI shall be governed according to:

  • Lawfulness;
  • Safety;
  • Proportionality;
  • Reliability;
  • Security;
  • Transparency;
  • Traceability;
  • Accountability;
  • Appropriate Explainability;
  • Privacy;
  • Fairness;
  • Non-Discrimination; And
  • Meaningful Human Oversight.

No AI functionality shall enter production solely because it is technically feasible.

ARTICLE 7 – PROHIBITED AI

Every AI use case shall undergo prohibited-practice screening.

FARENIO shall not knowingly place into service or use an AI practice prohibited by applicable law.

Potentially prohibited use cases shall be blocked pending Legal/Compliance determination.

Statutory prohibition overrides commercial objectives.

ARTICLE 8 – AI REGISTER

FARENIO shall maintain a controlled AI System Register.

Each material AI system shall record:

  • System ID;
  • Name;
  • Owner;
  • Technical Owner;
  • Provider;
  • FARENIO regulatory role;
  • Intended Purpose;
  • Affected Population;
  • AI Act classification;
  • Personal-Data Processing;
  • Automated-Decision Status;
  • Data Sources;
  • Model/Version;
  • Deployment Environment;
  • Human Oversight;
  • Transparency Requirements;
  • DPIA status;
  • Security Assessment;
  • Risk Assessment;
  • Residual Risk;
  • Monitoring;
  • Incidents;
  • Changes; And
  • Lifecycle Status.

ARTICLE 9 – AI LITERACY

Personnel operating or supervising AI shall receive AI literacy appropriate to:

  • Role;
  • Knowledge;
  • Experience;
  • System Complexity;
  • Deployment Context;
  • Affected Persons; And
  • Risk.

Enhanced training shall apply to personnel responsible for consequential AI systems.

ARTICLE 10 – AI LIFECYCLE

Material AI systems shall pass through:

Concept → Classification → Design → Privacy/Security Assessment → Development → Validation → Approval → Deployment → Monitoring → Change Control → Retirement.

Each stage shall produce documentation proportionate to risk.

ARTICLE 11 – HUMAN OVERSIGHT

Human oversight shall be meaningful rather than nominal.

Authorised personnel shall be capable, where appropriate, of:

  • Understanding System Limitations;
  • Reviewing Outputs;
  • Detecting Anomalies;
  • Disregarding Recommendations;
  • Overriding Outputs;
  • Intervening;
  • Escalating Concerns; And
  • Suspending Operation.

Personnel exercising oversight must possess sufficient authority and competence.

ARTICLE 12 – AUTOMATED DECISION-MAKING

Where AI processes personal data to make or materially support decisions affecting individuals, FARENIO shall assess GDPR automated-decision-making and profiling requirements.

FARENIO shall not make decisions based solely on automated processing producing legal or similarly significant effects unless legally permitted.

Required safeguards shall include, where applicable:

  • Transparency;
  • Human Intervention;
  • Opportunity To Express a Viewpoint;
  • Contestability; and
  • DPIA.

ARTICLE 13 – AI DATA GOVERNANCE

AI personal-data processing shall comply with:

  • Lawful-Basis Requirements;
  • Purpose Limitation;
  • Data Minimisation;
  • Accuracy;
  • Retention Limitation;
  • Confidentiality;
  • Access Control;
  • Provenance;
  • Data Quality;
  • Deletion Requirements;
  • International-Transfer Rules;
  • Data-Subject Rights; And
  • Privacy By Design/Default.

Access to Platform data does not automatically authorise use for AI training or fine-tuning.

ARTICLE 14 – FAIRNESS

Material AI shall be assessed, where relevant, for:

  • Representativeness;
  • Proxy Discrimination;
  • Discriminatory Correlations;
  • Differential Error Rates;
  • False Positives;
  • False Negatives; and
  • Disproportionate Effects.

AI shall never constitute an independent legal justification for unlawful discrimination.

ARTICLE 15 – AI TRANSPARENCY

Where legally required, individuals shall be informed when interacting with an AI system.

Applicable AI-generated or manipulated content shall carry legally required disclosure or labelling.

Automated-decision information required under data-protection law shall be provided through appropriate notices and interfaces.

ARTICLE 16 – THIRD-PARTY AI

Material AI suppliers shall undergo proportionate due diligence addressing:

  • Identity;
  • Regulatory Status;
  • Documentation;
  • Security;
  • Privacy;
  • Data Usage;
  • Subprocessors;
  • International Transfers;
  • IP;
  • Retention;
  • Incident Handling;
  • Auditability;
  • Continuity;
  • Model Changes; And
  • Exit Arrangements.

Supplier compliance representations do not replace FARENIO's own assessment.

ARTICLE 17 – AI VALIDATION AND MONITORING

Material AI shall be tested proportionately for:

  • Accuracy;
  • Reliability;
  • Robustness;
  • Security;
  • Bias;
  • Edge Cases;
  • False Positives/Negatives;
  • Integration;
  • Human Oversight;
  • Failure Behaviour;
  • Monitoring Readiness; And
  • Rollback Capability.

Production monitoring shall address:

  • Performance;
  • Drift;
  • Anomalies;
  • Misuse;
  • Security Events;
  • Complaints;
  • Discriminatory Effects; And
  • Regulatory Changes.

ARTICLE 18 – HIGH-RISK AI

Where a system qualifies as high-risk, FARENIO shall determine its regulatory role and implement the corresponding statutory obligations.

Controls shall include, where applicable:

  • Risk Management;
  • Data Governance;
  • Documentation;
  • Logging;
  • Transparency;
  • Instructions For Use;
  • Human Oversight;
  • Accuracy;
  • Robustness;
  • Cybersecurity;
  • Quality Management;
  • Conformity Assessment;
  • Registration;
  • Post-Market Monitoring; And
  • Incident Reporting.

ANNEX B

AI MANAGEMENT SYSTEM

The AI Management System shall contain:

Governance

  • Accountability;
  • Policies;
  • AI Ownership;
  • Regulatory Monitoring.

Planning

  • Classification;
  • Risk Assessment;
  • Stakeholder Assessment;
  • Compliance Planning.

Support

  • Competence;
  • AI literacy;
  • Training;
  • Documentation.

Operations

  • Inventory;
  • Procurement;
  • Development;
  • Validation;
  • Deployment;
  • Monitoring;
  • Incident Response;
  • Change Management;
  • Retirement.

Evaluation

  • KPIs;
  • Audits;
  • Incidents;
  • Management Review;
  • Corrective Action.

ANNEX C

AI SYSTEM REGISTER

Each material AI system shall record:

  • System ID;
  • System Name;
  • Business Owner;
  • Technical Owner;
  • Supplier;
  • Regulatory Role;
  • Intended Purpose;
  • Affected Population;
  • AI Act Classification;
  • Automated-Decision Classification;
  • Personal-Data Categories;
  • Legal Basis;
  • Data Sources;
  • Model;
  • Version;
  • Deployment Environment;
  • Human Oversight;
  • Transparency Requirement;
  • DPIA;
  • Security Assessment;
  • Risks;
  • Controls;
  • Residual Risk;
  • Review Frequency;
  • Incidents;
  • Changes; and
  • Lifecycle Status.

ANNEX D

AI DEPLOYMENT GATE

No material AI system shall enter production until FARENIO has determined, where applicable:

  1. Intended Purpose;
  2. Regulatory Role;
  3. Prohibited-Practice status;
  4. AI Act classification;
  5. Personal-Data Processing;
  6. Profiling/automated-decision status;
  7. DPIA requirement;
  8. Transparency Obligations;
  9. Human Oversight;
  10. Data-Governance Compliance;
  11. Security Validation;
  12. Model Validation;
  13. Supplier Due Diligence;
  14. Monitoring Readiness;
  15. Incident-Response Readiness; And
  16. Required Approval.

NO APPROVAL → NO PRODUCTION DEPLOYMENT.

ARTICLE 19 – RECORDS AND EVIDENCE

FARENIO shall maintain proportionate evidence demonstrating operation of this Policy and the wider FARENIO AI Governance & Compliance Framework, including, as relevant to AI Governance:

  • Policies;
  • Registers;
  • Risk Assessments;
  • AI Classifications;
  • DPIAs;
  • Security Assessments;
  • Verification Records;
  • Supplier Reviews;
  • Access Reviews;
  • Training Records;
  • Deployment Approvals;
  • Incidents;
  • Complaints;
  • Corrective Actions;
  • Audits; And
  • Management Reviews.

ARTICLE 20 – LEGAL PRECEDENCE

This Policy forms part of the internal governance architecture of GQ VENTURES SAS operating under the FARENIO trade name. It shall operate together with the FARENIO Terms and Conditions, Service Provider Terms, Vendor/Partner Terms, Privacy Policy, Cookie Policy, Consumer Information, Information Security Policies, Incident Response Plan, Data Retention Policy, Business Continuity Plan, contractual payment-provider arrangements and applicable tax and accounting procedures.

Nothing in this Policy limits mandatory statutory rights or obligations.

Where a conflict exists between this Policy and mandatory French or European Union law, mandatory law prevails.

ARTICLE 21 – CORE GOVERNANCE RULE

FARENIO adopts the following control principle:

VERIFY PEOPLE → PROTECT USERS → PROTECT DATA → SECURE SYSTEMS → GOVERN AI → CONTROL PAYMENTS → REPORT TAX DATA → MONITOR → AUDIT → IMPROVE.

No commercial objective, automated system, Service Provider, employee, contractor, Vendor, Partner or third-party technology provider may override a mandatory legal or regulatory requirement.

ARTICLE 22 – CONTACT

AI matters:

AI Governance Function

Email: contact@farenio.com

Privacy requests shall be submitted through the privacy contact identified in the FARENIO Privacy Policy.

Security incidents shall be submitted through FARENIO's designated security-reporting channel.

Loading...