1. Document Control
| Field | Value |
|---|---|
| Document | YourTransfer24™ Responsible AI Use Policy |
| Policy ID | YT24-POL-RESPONSIBLE-AI |
| Version | 1.0.0 |
| Status | ACTIVE |
| Classification | PUBLIC |
| Effective date | 3 October 2026 |
| Policy Owner | YourTransfer24™ |
| Applies to | Employees, contractors, developers, administrators and other authorized users of AI in connection with YT24 work |
| Review cycle | Periodic, and upon a material legal, regulatory, security or operational change |
| Canonical language | English |
This version is in force. It is the Responsible AI Use Policy of YourTransfer24™. It replaces any shorter responsible-AI statement previously published for this Policy ID.
2. Purpose
This Policy protects customers, travelers, providers, drivers, the workforce, YourTransfer24, confidential information, official records, technical information, credentials, security and intellectual property from inappropriate use of artificial intelligence.
AI may assist YT24 work. It does not become the owner of a fact, the authority for a transaction, or a substitute for the user's responsibility.
3. Scope
This Policy applies to every employee, contractor, developer, administrator and other authorized user who uses AI in connection with YT24 work, wherever that user works.
It covers external AI tools, AI embedded in a YourTransfer24 system, coding assistants, generative AI, automation and agents, and the model providers behind those systems. It also covers any use of AI in an employment decision.
It supplements the Internal Use Terms & Employee Responsibilities, the Customer Information Handling & Confidentiality Policy, the Security & Access Control Policy and the Security & Incident Escalation Policy. Where one of those documents, a contract, or applicable law sets a stricter rule, the stricter rule applies.
4. Definitions
An AI system is a system that infers, generates, classifies, recommends or acts from data, including a generative model, an agent, a coding assistant and an embedded model. A deterministic rules engine that does not infer or generate in that way is not, by itself, an AI system under this Policy.
Approved AI is an AI service, account, integration, tool and data scope that YourTransfer24 has authorized for a stated use. A consumer or personal AI account is not approved, even when the underlying model is the same model an approved service uses.
An official record is a current business fact held by the authoritative YourTransfer24 system for that fact, such as a booking state, a price, an availability, a payment state, or an assignment. Approved knowledge is versioned knowledge an authorized person has approved for a defined scope. A candidate insight is a proposal drawn from patterns. It is not knowledge and it is not a fact. A conversation is evidence that something was said. It is not evidence that the statement is true.
Material output is AI output that could affect a customer, a traveler, a provider, a driver, a worker, a payment, a security control, a legal or compliance position, or a canonical record.
5. YT24 AI Governance Principles
AI is not transactional authority.
This Policy establishes YT24's minimum global standard for responsible use of artificial intelligence. Stricter applicable legal, regulatory, employment, privacy or security requirements apply where required.
AI outputs do not override official records, authoritative systems, approved policies, or decisions reserved to authorized people or governed processes. Approved knowledge overrides conversation context. Conversation context may help interpret what a person is asking. It does not override an official record.
External model providers supply reasoning and generation. They do not own YourTransfer24 business records, approved knowledge, or official records. Agents act only within capabilities and permissions YourTransfer24 has governed.
YourTransfer24 may automate a repeatable action. Automation still requires authorization from the system or process that owns the record. Confidence is not authority. An AI answer does not become a booking, a price, a payment, or an assignment because the model stated it.
6. Approved AI Tools
YT24 work that involves confidential, personal, operational or proprietary information is done only through an AI service YourTransfer24 has approved for that information and that use.
The user does not decide which model or which provider may receive YT24 information. Approval of a tool for one task is not approval for every task. Approval of a tool for one data class is not approval for a more sensitive class.
The user uses the approved account and the approved integration. The user does not point an approved tool at a personal account, a personal API key, or a side subscription.
8. Customer Information
The user must not submit identifiable customer or traveler information to an unapproved AI system. That includes identity data, contact data, booking data, itinerary data, location data, and any document that contains those details.
Approved YT24 AI may process customer or traveler information only inside the scope YourTransfer24 has authorized for that workflow. The user must not widen that scope by pasting additional records, by connecting an export, or by asking the tool to retrieve a customer the task does not concern.
Output that contains customer information is handled under the Customer Information Handling & Confidentiality Policy. The user must not copy it into an unapproved tool, an unapproved chat, or a personal account.
9. Workforce Information
The same limit applies to workforce personal information. The user must not submit protected information about an employee, a contractor, an applicant or another authorized user to an AI system that is not authorized for that information.
Protected workforce information includes identity and contact details, employment and performance records, compensation, medical or accommodation information, disciplinary material, access records, and the contents of a personnel file. A business title and a work task that contain none of that information are not, by themselves, a personnel file.
The user must not ask an AI system to infer a worker's health, disability, religion, union activity, or other sensitive characteristic, or to score workers against one another for a purpose section 14 regulates.
10. Confidential / Proprietary Information
The user must not disclose the following to an AI system except through a service YourTransfer24 has approved for that exact class of information:
- a password, an API key, a token, a recovery code, or an environment secret;
- security architecture or a vulnerability;
- unpublished commercial data;
- confidential information of a provider or another partner;
- a confidential contract;
- proprietary source code, where the tool is not approved for code; and
- incident evidence.
A coding assistant may be used on YourTransfer24 code only when YourTransfer24 has approved that assistant for that work. Approval to explain a short excerpt is not approval to upload source code, a secret, or a configuration.
The user must not paste incident evidence, a log that contains personal data, or a credential into an AI system in order to interpret an incident, unless that system is the approved tool for that incident work.
11. Verification Obligation
The user verifies material AI output before relying on it. Verification uses the canonical source, not a second prompt that repeats the same answer.
The user checks names, dates, places, prices, booking references, policy statements, translations that will reach a customer or a partner, and any instruction the output gives about a security control or a legal duty.
If the output cannot be verified, the user does not send it, does not enter it in a record, and does not treat it as a fact. Fluency, confidence, and a citation the user has not opened are not verification.
13. Knowledge Governance
A conversation with a customer, a provider, a driver, or a worker must not automatically become a canonical fact, modify this Policy or any other policy, modify approved knowledge, retrain a YT24 system, or create a durable business rule.
AI may identify a candidate insight, including a repeated question, a knowledge gap, or a possible correction. A candidate insight is not used to answer the next person, and is not written into approved knowledge, until an authorized person approves it for a defined scope. The AI system does not approve its own candidate. How often a statement appears is not proof that the statement is true.
Approved AI uses approved knowledge and official records. It does not use an unapproved candidate as if it were approved. Approved knowledge still does not override an official record.
14. Employment Decisions
Unapproved AI must not make, or materially determine, a decision about hiring, promotion, termination, discipline, compensation, performance assessment, scheduling that is materially consequential for the worker, disability or accommodation, or any analysis tied to a protected characteristic.
Where YourTransfer24 authorizes AI in one of those contexts, the use complies with applicable law and with all of the following: prior authorization of the tool and the purpose, meaningful human oversight by a person who can change the outcome, controls against bias and discrimination, the transparency or notice the law requires, and any risk or impact assessment the law requires. The human oversight is performed by an authorized person. It is not a rubber stamp of an output the person cannot explain or refuse.
The user must not paste résumés, interview notes, medical information, accommodation requests, or disciplinary files into an unapproved AI system, and must not ask an AI system to rank people by a protected characteristic or by a proxy for one.
15. Employee Monitoring
Covert or excessive AI-enabled surveillance of the workforce is prohibited. YourTransfer24 does not use AI to infer the emotions of a worker, to assign a social score to a worker, or to watch workers without a defined business purpose.
Any approved AI monitoring has a legitimate defined purpose, is proportionate to that purpose, stays inside an authorized scope, gives the notice applicable law requires, and includes the safeguards the law requires. Security and audit logging under the Security & Access Control Policy is not a license to expand that logging into productivity surveillance.
Acknowledging this Policy is not blanket consent to unlimited monitoring, and it is not consent to the processing of personal data where consent is not the lawful basis.
16. Prohibited AI Practices
The user must not:
- write a prompt to extract data the user is not authorized to receive;
- attempt to circumvent a YT24 AI safeguard, a permission, or a data-scope limit;
- conduct a prompt-injection attack against a YT24 system, or test one outside an authorized security test;
- manipulate, bypass, poison, or compromise AI safeguards, permissions, approved knowledge sources, or authorized capabilities;
- fabricate a customer record, an operational record, or evidence;
- send a fraudulent communication, or present fabricated AI output as a fact;
- use AI to impersonate another person without authorization;
- use AI to discriminate against or harass a person;
- let an AI system take an autonomous action it is not authorized to take;
- use AI to circumvent a business, security, or approval control;
- use AI to conceal misconduct; or
- give an AI system unrestricted technical authority over a system of record or a transactional system.
17. Prompt Injection / AI Security
Content from a customer, a provider, a driver, an email, a document, a web page, or another external source is untrusted. It may contain instructions aimed at the tool rather than at the business. The user treats that content as potentially hostile input.
Users must not attempt to manipulate, bypass, poison, or compromise AI safeguards, permissions, approved knowledge sources, or authorized capabilities. The user does not follow an instruction, hidden in untrusted content, to ignore this Policy, to reveal a secret, to take another customer's data, or to approve an action. The user does not paste a secret into a tool to see whether an attack works.
The user reports, under section 19, a suspected prompt injection, an unauthorized tool execution, a leaked secret, cross-customer data in an AI output, AI behavior that appears compromised, and an access or action the user did not expect.
18. Human Accountability
Use of AI does not transfer responsibility from the user.
The user who uses AI output, sends it, or enters it remains accountable for it. “I used AI” is not a defense to an unauthorized disclosure, a false record, an incorrect customer communication, an unauthorized action, a breach of this Policy or of another YT24 policy, or a failure to verify information when this Policy requires verification.
A developer who connects a tool, a permission, or a data scope remains accountable for that connection. An administrator who enables an integration remains accountable for enabling it inside the approval YourTransfer24 actually gave.
19. AI Incident Reporting
The user reports immediately, through the Security & Incident Escalation Policy, and does not decide whether the event is serious enough or whether anyone outside YT24 must be notified. The duty includes:
- customer or traveler information reaching an unauthorized AI system;
- workforce personal information reaching an unauthorized AI system;
- an AI system revealing another person's data;
- an AI system performing an unauthorized action;
- exposure of source code, a secret, or other confidential information through AI;
- a prompt injection that may have succeeded;
- a material operational error caused by AI; and
- AI security that appears compromised.
The user stops the further use that is still exposing the information, where it is safe to do so, preserves the evidence, and does not delete the prompt or the output to hide the event. The user does not ask the same unapproved tool to forget the data as a substitute for the report.
20. AI Literacy
The user completes the AI training YourTransfer24 assigns for the user's role before using AI in that role, and again when the assignment changes. The training covers the limits of AI, hallucination, privacy, customer confidentiality, security, prompt injection, canonical authority, verification, permitted and prohibited uses, and incident escalation.
A person who builds, configures, or supervises an agent or an integration receives the additional training that role requires, including tool permissions and the rule that the model does not own the business fact.
Where Article 4 of Regulation (EU) 2024/1689 applies, that training is how YourTransfer24 gives the person a level of AI literacy sufficient for the person's knowledge, experience, and the context in which the AI system is used. Completing a course does not authorize a use this Policy prohibits.
21. Model / Provider Governance
YT24 governance determines which models, vendors, accounts, integrations, tools, and data scopes are approved. The user cannot approve an external provider, extend an approval, or accept a provider's terms for YT24 information.
YourTransfer24 keeps provider independence. A model provider is replaceable. No provider is the system of record, and no workflow is designed so that YT24 business truth exists only inside that provider.
A change of provider, a new tool, or a wider data scope is not in force because a user has tested it. It is in force when the governed approval says so.
22. Auditability
Use of approved AI that touches protected information, a permission, or an official record is traceable to the workflow, the user or service identity, and the time, to the extent YourTransfer24's approved controls provide. The user must not disable that trace or move the work to an unlogged personal account to avoid it.
Those records support security, confidentiality, fraud prevention, incident detection, and compliance. They are not a mandate for unlimited monitoring of the workforce, and acknowledgement of this Policy is not consent to that monitoring.
23. Enforcement
A breach of this Policy may result, as applicable law and the user's contract allow, in restriction or suspension of access, investigation, required retraining, other disciplinary action, and termination of employment or of the engagement. YourTransfer24 may preserve evidence, seek civil remedies available to it, and refer a matter to a public authority where that is appropriate.
Putting protected information into an unapproved AI system, fabricating a record, circumventing a safeguard, and using AI to conceal misconduct are serious misconduct. A good-faith mistake that the user reports promptly is handled under the Security & Incident Escalation Policy and is not serious misconduct merely because the user reported it.
This Policy does not require the user to indemnify YourTransfer24. It does not authorize a deduction from wages or other pay that applicable law forbids. It does not create an automatic personal debt, fine or penalty. Any consequence is decided under applicable law and the terms of the user's employment or engagement.
24. Jurisdictional Application
This Policy is the global baseline for every authorized user, including users who work in or access YT24 systems from the Dominican Republic, the United States, Canada, the European Economic Area, the United Kingdom, Brazil, Singapore, Australia, or any other place where YourTransfer24 authorizes access.
The user's location does not lower the ban on unapproved AI, and it does not allow customer information to be handled below the Customer Information Handling & Confidentiality Policy. Each statute below is applied only where it covers the event, on its own date and its own definitions. A rule from one jurisdiction is not copied onto another.
24.1 European Union
Where Regulation (EU) 2024/1689 applies, YourTransfer24 does not place on the market, put into service, or use a practice that Article 5 prohibits on the date that prohibition applies. Prohibited practices that have applied since 2 February 2025 include harmful manipulative or deceptive techniques, exploitation of vulnerabilities, social scoring, untargeted scraping of facial images to build a recognition database, inferring sensitive attributes by biometric categorization, and emotion recognition in the workplace except for a narrow medical or safety use the Regulation itself allows. YourTransfer24 does not operate that workplace exception. Further prohibitions are applied from the date the Regulation gives them, including 2 December 2026 where that is the date set.
Article 4 has applied since 2 February 2025. Section 20 is the workforce rule that meets it. Transparency duties in Article 50 that have applied since 2 August 2026, including a duty to make clear that a person is interacting with an AI system or to mark synthetic content, are carried out in the approved workflow. The user must not remove that disclosure.
Employment and worker-management uses are capable of being high-risk under Annex III. The high-risk obligations for Annex III systems apply from 2 December 2027 where the Regulation sets that date. Before and after that date, section 14 still forbids an unapproved tool from making or materially determining an employment decision, and still requires human oversight, non-discrimination, required notice, and any assessment the law already imposes. Those later statutory duties are added when they apply. They are not imposed on a country the Regulation does not cover.
Where the GDPR applies, personal data placed in an AI system is processing and is subject to the GDPR's principles, security duty, and transparency duty. A decision based solely on automated processing, including profiling, that produces a legal or similarly significant effect is not made that way for a section 14 decision. The central process supplies the human intervention and the information Articles 13, 14, 15 and 22 require. The user does not choose the legal basis and does not create an international-transfer mechanism by opening an account.
24.2 United Kingdom
Where UK data protection law applies, personal data used in an AI system remains subject to that law, including fairness, accuracy, security, and transparency. Where that law gives a person a right to human intervention, to express a view, or to contest a significant decision based on automated processing, the authorized process honors that right. The Information Commissioner's guidance on AI and on explaining AI-assisted decisions is followed in that design. It does not authorize an employee to use an unapproved tool.
An employment decision assisted by AI remains subject to the Equality Act 2010. A tool must not discriminate because of a protected characteristic. Meaningful human oversight under section 14 is required even when a statute's automated-decision provision is not the provision that applies.
24.3 United States
There is no single United States statute that sets one nationwide process for AI in employment. Title VII of the Civil Rights Act, the Americans with Disabilities Act, and the Age Discrimination in Employment Act apply to an employment decision that software or AI assists, where those statutes apply. YourTransfer24 remains responsible for a discriminatory outcome of a tool a vendor supplied. A selection procedure that has an adverse impact must be job-related and consistent with business necessity where Title VII so requires. A tool must not unlawfully screen out a person with a disability or elicit disability-related information.
A state or local rule is applied on its own geography, its own definition of a covered tool, and its own effective date. Where New York City Local Law 144 applies to an automated employment decision tool, the tool is not used unless the required bias audit and the required notices are in place. A California employment rule on automated-decision systems is applied where the Fair Employment and Housing Act and that rule cover the practice. A statute that is not yet in force is not applied early, and a statute that covers one city or one data practice is not treated as national.
The Federal Trade Commission Act prohibits unfair or deceptive practices. The user must not use AI to mislead a customer, a traveler, a provider, or a worker, and must not present an AI-generated claim YourTransfer24 cannot substantiate.
24.4 Canada
Canada has no general artificial-intelligence statute in force that replaces PIPEDA. Where PIPEDA or a provincial privacy statute applies, personal information placed in an AI system is a collection, use, or disclosure that must meet that statute. The user does not make that determination and does not open a personal account to avoid it.
Employment decisions remain subject to applicable human-rights law. Section 14 applies whether or not a provincial privacy statute also regulates automated decisions. A federal determination does not replace a provincial duty that has its own scope.
24.5 Dominican Republic
Ley 172-13 requires security measures and a duty of secrecy for personal data. This Policy does not treat a separate Dominican artificial-intelligence statute as in force. Putting personal data into an unapproved AI system is a breach of this Policy and can breach that duty of secrecy. The GDPR timetable is not imported into the Dominican Republic by default.
24.6 Brazil
Where the LGPD applies, personal data used in an AI system is processing under that law, including the principles of purpose, necessity, non-discrimination, transparency, and security. Article 20 gives the data subject a right to request review of a decision based solely on automated processing of personal data that affects the data subject's interests, and a right to clear information about the criteria and procedures used, subject to the limits in that article. The central process answers that request. A bill that has not become law is not applied as if it had.
24.7 Singapore
Where the PDPA applies, personal data used to develop or to operate an AI recommendation or decision system remains subject to the PDPA, including consent or another lawful basis, notification, purpose limitation, and protection. The user does not decide that a business-improvement exception applies. The PDPC's advisory guidelines are used by the central process. They do not authorize a personal account.
24.8 Australia
Where the Privacy Act and the Australian Privacy Principles apply, personal information entered into an AI product, and personal information in the output, is subject to those principles, including accuracy, security, and the notice the Act requires. A use that has a legal or similarly significant effect on a person is a high privacy risk and is not given to an unapproved product. The employee-records exemption does not cover customer or traveler information, and it does not cover a contractor to whom the exemption does not apply.
24.9 NIST frameworks
YourTransfer24 uses the NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0) and the Generative Artificial Intelligence Profile (NIST AI 600-1) as the voluntary method for governing, mapping, measuring, and managing AI risk, including generative-AI risk. Those documents do not replace a statute and do not authorize a use this Policy prohibits.
25. Severability
Each provision of this Policy is applied to the maximum extent the law allows. If a provision cannot be applied to a particular user, a particular place, or a particular set of facts, it is applied as far as it lawfully can be, and the rest of the Policy continues in force. A provision that cannot be applied at all in that situation is severed for that situation only.
No user may read a limit that the law imposes on one provision as permission to ignore a duty that remains lawful.
26. Electronic Acknowledgement
Acknowledgement of this Policy is a condition of authorized access to YT24 protected internal systems, subject to applicable law. This Policy is an active mandatory corporate policy. It is not one of the three acknowledgements presented each time the user signs in, and completing those three acknowledgements does not satisfy this Policy.
The authoritative record is the formal versioned acknowledgement. When a user first receives authorized access, and whenever YourTransfer24 publishes a new version of this Policy, the user must complete that acknowledgement before protected access is granted or restored.
YourTransfer24 presents the applicable version to the authenticated user. The user must be able to read the complete Policy and must reach the end of the presented Policy before the acknowledgement action becomes available. The user then enters their full name and affirmatively acknowledges the Policy.
The acknowledgement date and time are generated by YourTransfer24 using an authoritative server timestamp. The user does not manually enter, backdate, or future-date that time.
By completing the formal acknowledgement, the user confirms that the Policy was presented, that the user had the opportunity to review it, and that the user understands the obligation to comply with it as a condition of authorized access to YT24 systems, subject to applicable law.
Electronic acknowledgement of this Policy is not, by itself, consent to the processing of personal data where consent is not the applicable lawful basis for that processing. It is not consent to unlimited monitoring.
YourTransfer24™ maintains versioned and auditable records of required policy acknowledgements.
Additional technical evidence may be retained where authorized, proportionate, and permitted by applicable law and workforce privacy requirements.
An acknowledgement applies only to the exact version recorded. A new version requires a new formal acknowledgement. Previous versions and previous acknowledgements stay in the compliance record and are not silently overwritten.
The recurring sign-in acknowledgements for the other workforce policies remain separate. They must not be used to fabricate or infer acknowledgement of this Policy or of a specific version of it.
26.1 Relationship Between Authentication, Policy Compliance and Authorization
Authentication, policy compliance and application authorization are separate controls.
The governing sequence is:
Authentication→ Policy Compliance→ Application Authorization→ Access
Successful authentication does not bypass an outstanding formal acknowledgement of this Policy.
Formal acknowledgement of this Policy does not independently grant access to an application for which the user is not otherwise authorized.
Recurring sign-in acknowledgement does not replace either formal policy compliance or application authorization.

Issued by YourTransfer24 for internal use. Where your employment agreement or the internal compliance programme sets stricter terms, those terms apply.