Responsible Artificial Intelligence Policy
Version 1.0 · Last updated
Passages marked [POR DEFINIR] await the company details still to be confirmed.
English translation for convenience — the Spanish version is the governing text.
TITLE I. GENERAL PROVISIONS
Chapter 1. Purpose and Scope of Application
1.1 Purpose
This Responsible Artificial Intelligence Policy (hereinafter, the "Policy") is intended to establish the guiding principles, governance mechanisms, and the specific disclaimer regime applicable to the use of Artificial Intelligence technologies within the Ancuria™ Inteligencia Inmobiliaria platform (hereinafter, "Ancuria", the "Platform" or the "Data Controller"), supplementing, on this specific subject matter, the provisions of the Terms and Conditions of Service and of the Full Privacy Notice.
1.2 Legal nature
This Policy constitutes both an internal governance document, insofar as it establishes the principles governing the development and operation of Ancuria's Artificial Intelligence systems, and a contractual instrument of information and disclaimer toward the User, insofar as it specifies the scope, limitations and liability regime applicable to the use of such technology.
1.3 Relationship to other documents
This Policy must be interpreted jointly with the Terms and Conditions of Service, particularly its Title V (Intellectual Property, regarding Chapters 14 and 18) and its Title IX (Limitation of Liability, particularly Chapter 40), as well as with the Full Privacy Notice, particularly its Chapter 9. In case of discrepancy regarding the specific subject matter of Artificial Intelligence, the provisions of this Policy shall prevail, as it is the instrument specialized in that subject matter.
1.4 Material scope of application
This Policy applies to every Artificial Intelligence component integrated into the Platform, including, without limitation: the conversational Artificial Intelligence Agents; the recommendation engine; the models used to generate Market Indicators, without prejudice to those Indicators being calculated by the deterministic Calculation Engine described in section 1.6; and any other component the Platform incorporates in the future whose operation rests, wholly or partially, on machine learning or natural language processing models.
1.5 Personal scope of application
This Policy is applicable to Ancuria in its capacity as developer and operator of those components, to the Users (Real-Estate Agents, Real-Estate Developers and Companies) who use those components in the exercise of their professional activity, and to the end clients who interact with the Artificial Intelligence Agents in the context of that activity.
1.6 Fundamental distinction: Calculation Engine vs. generative Artificial Intelligence
For purposes of this Policy it is essential to distinguish between: (i) the Calculation Engine, a deterministic component executing all of the Platform's financial and mathematical operations (mortgage calculations, return on investment, capital appreciation, payment plans) according to fixed, auditable formulas, with no margin of variability or interpretation; and (ii) the generative or conversational Artificial Intelligence Agents, which interpret requests and generate responses in natural language through probabilistic models. The disclaimer provisions of this Policy regarding imprecision or variability of results apply to category (ii); the Calculation Engine, by its deterministic nature, is not subject to such variability, without prejudice to both components generating informational results pursuant to Title IV.
1.7 Consequences of non-compliance
The Data Controller's non-compliance with the governance provisions contained in this Policy may give rise to liability under applicable legislation; the User's non-compliance with the obligations applicable to them under Title IV may give rise to the consequences provided in the Terms and Conditions of Service.
Chapter 2. Definitions
2.1 "Artificial Intelligence"
For purposes of this Policy, Artificial Intelligence is understood as the set of computational systems capable of performing tasks that, if executed by an individual, would require a degree of intelligence, including natural language processing, machine learning, and recommendation systems, integrated into the Platform under the terms of Chapter 1.
2.2 "Artificial Intelligence Agent" or "AI Agent"
A software component based on generative Artificial Intelligence models, integrated into the Platform, capable of sustaining conversational interactions, interpreting requests in natural language, and generating responses or executing automations within the parameters defined by the Data Controller.
2.3 "Language model" or "LLM"
An Artificial Intelligence model trained to process and generate text in natural language, whether developed by the Data Controller or licensed from a third-party Artificial Intelligence Provider pursuant to section 2.7.
2.4 "Calculation Engine"
The deterministic technology component described in section 1.6, expressly differentiated from the Artificial Intelligence Agents for purposes of this Policy.
2.5 "Hallucination"
A technical phenomenon inherent to generative language models, consisting of the generation of incorrect information, inconsistent with reality, or unsupported by the source data, presented by the model with an appearance of certainty.
2.6 "Algorithmic bias"
The systematic tendency of an Artificial Intelligence model to produce results that unjustifiably favor or harm a particular group or category of persons, derived from the training data, the model's design, or its implementation.
2.7 "Third-party Artificial Intelligence Provider"
A legal entity other than the Data Controller that provides the technology infrastructure or underlying language models used by the Platform's Artificial Intelligence Agents, acting pursuant to the provisions of Chapter 16.
2.8 "Human oversight" or "human-in-the-loop"
A governance mechanism through which an individual reviews, validates, corrects or intervenes in the operation or results of an Artificial Intelligence system, in the terms described in Chapter 5.
2.9 "Automated decision with definitive legal effects"
A decision adopted exclusively on the basis of automated data processing, without human intervention, that produces legal effects on a person or significantly affects them, a category from which the results of the Platform's Artificial Intelligence Agents and Calculation Engine are expressly excluded, pursuant to Chapter 12.
2.10 "Artificial Intelligence Incident"
Any event in which an Artificial Intelligence component of the Platform generates a manifestly erroneous, discriminatory, unsafe result, or one that substantially departs from expected behavior, as detailed in Chapter 14.
Chapter 3. Guiding Principles of Responsible Artificial Intelligence
3.1 Legal objective
To establish the principles governing the design, development, implementation and operation of the Platform's Artificial Intelligence components, as a reference framework for Ancuria's governance decisions and as an interpretive parameter of this Policy.
3.2 Principle of transparency
The Data Controller undertakes to clearly and accessibly inform when a User or end client is interacting with an Artificial Intelligence Agent, in the terms developed in Chapter 7, avoiding any practice that induces the belief that such interaction is with an individual when it is not.
3.3 Principle of human oversight
The Data Controller designs its Artificial Intelligence components so that there exists, to the extent the nature of each feature permits, a human oversight mechanism, whether through prior or subsequent review, or through the possibility of human intervention upon a request or result that warrants it, pursuant to Chapter 5.
3.4 Principle of non-discrimination and fairness
The Data Controller adopts reasonable measures to prevent, detect and correct algorithmic biases that could generate discriminatory treatment toward any person based on categories protected by applicable legislation, pursuant to Chapter 8.
3.5 Principle of security and robustness
The Data Controller implements reasonable technical controls to prevent misuse of its Artificial Intelligence components, including manipulation attempts, extraction of confidential information, or generation of harmful content, pursuant to Chapter 9.
3.6 Principle of accountability
The Data Controller maintains internal governance, documentation and traceability mechanisms regarding the operation of its Artificial Intelligence components, allowing it to handle audits, Due Diligence processes, and requests from competent authorities.
3.7 Principle of informational limitation
The Data Controller designs its Artificial Intelligence components so that their results have the character of informational support for the User's professional decision, and not that of a substitute for such professional judgment or for specialized advice, as extensively developed in Title IV.
3.8 Principle of continuous improvement
The Data Controller periodically reviews and updates its Artificial Intelligence components, incorporating User feedback, the findings of its oversight processes, and the evolution of international best practices in the field, pursuant to Chapter 15.
3.9 Nature of the principles
The principles set out in this Chapter constitute governance commitments and interpretive parameters of this Policy; their specific operational development, as well as the corresponding liability regime, are established in the subsequent Titles.
TITLE II. GOVERNANCE AND HUMAN OVERSIGHT
Chapter 4. Artificial Intelligence Governance Structure
4.1 Legal objective
To describe the internal structure through which the Data Controller administers the risks associated with the development and operation of its Artificial Intelligence components, in compliance with the accountability principle indicated in section 3.6.
4.2 Artificial Intelligence governance officer
According to its organizational structure, the Data Controller will designate a person or team responsible for overseeing compliance with this Policy, whose identity may be made known to auditors, investors or competent authorities that so require for Due Diligence purposes, without such formal designation, as of the date of this document, being a requirement for the validity of this Policy.
4.3 Internal technical documentation
The Data Controller will maintain internal technical documentation regarding each Artificial Intelligence component in operation, including, to the reasonable extent, the description of its functionality, the third-party Artificial Intelligence Providers involved, and the applicable human oversight mechanisms, which may be made available in audit or Due Diligence processes under the corresponding confidentiality agreements.
4.4 Prior assessment of new components
Prior to the implementation of a new Artificial Intelligence component or a material modification of an existing one, the Data Controller will conduct a reasonable assessment of the associated risks, including security risks, algorithmic bias risks, and risks of generation of inaccurate results, documenting the mitigation measures adopted.
4.5 Consequences of non-compliance
The lack of reasonable documentation or assessment under this Chapter does not, by itself, generate liability toward individual Users, without prejudice to it possibly being considered a relevant factor in determining the Data Controller's diligence in a claim based on an Artificial Intelligence Incident.
Chapter 5. Human Oversight
5.1 Legal objective
To detail the mechanisms through which the Data Controller incorporates human intervention and oversight into the operation of its Artificial Intelligence components, developing the principle indicated in section 3.3.
5.2 Ex ante oversight through Agent design
The Data Controller configures the Artificial Intelligence Agents with instructions and operational limits defined by its technical team, restricting the scope of their responses to the previously approved features and tone, as the first mechanism of human control over the system's behavior.
5.3 Concurrent oversight through monitoring
The Data Controller implements reasonable monitoring mechanisms over the operation of the Artificial Intelligence Agents in production, enabling detection of anomalous behavior patterns or repeated generation of imprecise results.
5.4 Ex post oversight through report review
The Data Controller reviews the reports of erroneous results that Users submit pursuant to the mechanism described in Chapter 14, incorporating such feedback into its continuous improvement processes.
5.5 Human intervention at the User's request
When the nature of a matter warrants it, particularly where an end client expresses dissatisfaction with an interaction held with an Artificial Intelligence Agent, the corresponding Real-Estate Agent may — and in certain cases must, according to their own professional judgment — personally intervene in the matter, without the Platform limiting that possibility of human intervention.
5.6 Limits of real-time human oversight
The User acknowledges that, given the real-time nature of the Artificial Intelligence Agents' operation, it is not technically feasible for every individual generated response to be reviewed by an individual prior to delivery, so the human oversight described in this Chapter operates primarily through the design, monitoring and subsequent review mechanisms indicated in sections 5.2 through 5.4, without prejudice to the possibility of direct intervention indicated in section 5.5.
5.7 Consequences of non-compliance
The absence of the reasonably required human oversight mechanisms under this Chapter may be considered a relevant factor in determining the Data Controller's diligence in a claim based on an Artificial Intelligence Incident, without this modifying the disclaimer regime established in Title IV regarding the informational nature of the generated results.
Chapter 6. Model Lifecycle and Risk Management
6.1 Legal objective
To describe the Data Controller's general approach to managing the risk associated with the lifecycle of the Artificial Intelligence models integrated into the Platform.
6.2 Selection of third-party Artificial Intelligence Providers
The Data Controller selects its third-party Artificial Intelligence Providers considering, among other factors, their industry track record, their public security and responsible-use policies, and the contractual guarantees they offer regarding availability, confidentiality and quality of service.
6.3 Pre-implementation testing
Before putting a new Artificial Intelligence component into production, the Data Controller conducts reasonable tests aimed at verifying its operation according to the intended purpose, without such tests constituting a guarantee of total absence of errors, given the probabilistic nature of this technology, expressly recognized in Chapter 10.
6.4 Continuous post-implementation monitoring
Once in production, the Data Controller maintains continuous monitoring mechanisms over the performance of its Artificial Intelligence components, pursuant to section 5.3, enabling timely identification of degradation in the quality of the generated results.
6.5 Updating and retraining
The Data Controller may update, retrain or replace the underlying models of its Artificial Intelligence components as the available technology evolves, without such update, by itself, requiring a material amendment of this Policy, unless it materially alters the guiding principles indicated in Chapter 3.
6.6 Consequences of non-compliance
The provisions of this Chapter describe the Data Controller's risk-management approach for transparency purposes toward Users, investors and auditors, without the description of that approach generating specific result obligations other than those expressly agreed in the Terms and Conditions of Service.
TITLE III. TRANSPARENCY AND RESPONSIBLE USE
Chapter 7. Transparency Toward the User and the End Client
7.1 Legal objective
To develop the transparency principle indicated in section 3.2, establishing the concrete mechanisms through which the Data Controller informs of the artificial nature of its Artificial Intelligence components.
7.2 Identification of the Artificial Intelligence Agents
Every Artificial Intelligence Agent that interacts directly with a User or an end client within the Platform will expressly identify itself as such at the start of the interaction, through a clear statement (for example, "You are interacting with an Ancuria Artificial Intelligence assistant"), and in no case will the Data Controller design its Agents to pass themselves off as an individual.
7.3 Obligation of the Real-Estate Agent toward their client
When a Real-Estate Agent enables their own clients' interaction with a Platform Artificial Intelligence Agent (for example, in the autonomous customer-service module), that Real-Estate Agent undertakes not to remove, hide or obscure the identification indicated in section 7.2, and to additionally inform their client, according to their own professional judgment, about the use of such technology in their service process.
7.4 Explainability of results
The Data Controller seeks to have the Artificial Intelligence Agents and Market Indicators present their results in a way understandable to a non-technical user, avoiding, as far as possible, presenting conclusions without the context or factors supporting them.
7.5 Communication of limitations
The Data Controller will incorporate, within the Platform's own interface, reminder notices about the informational and non-binding nature of the results generated by the Artificial Intelligence Agents and the Market Indicators, at the points of the Platform where such reminder is pertinent according to the Data Controller's design criteria.
7.6 Consequences of non-compliance
The deliberate design of an Artificial Intelligence Agent to hide its artificial nature from the User or the end client shall constitute a material breach of this Policy, and may give rise to additional liability under applicable consumer protection legislation.
Chapter 8. Non-Discrimination and Algorithmic Fairness
8.1 Legal objective
To develop the non-discrimination principle indicated in section 3.4, in the specific context of a real-estate platform, a sector in which algorithmic fairness carries particular social and international regulatory relevance.
8.2 Prohibition of discriminatory criteria in design
The Data Controller does not design its Artificial Intelligence components to consider, as a recommendation, rating or prioritization factor, categories protected by applicable legislation, such as ethnic origin, gender, religion, sexual preference, or disability, regarding the individuals involved in a real-estate transaction.
8.3 Review of training data and market sources
The Data Controller, to the extent of its technical capabilities and the nature of the available market data, makes reasonable efforts to identify and mitigate biases that could be present in the information sources used to generate Market Indicators, acknowledging that such sources may reflect historical patterns of inequality in the real-estate market foreign to the Platform's design.
8.4 Prohibition of discriminatory use by the User
It is expressly prohibited for the User to instruct, through prompts, configurations or any other mechanism, a Platform Artificial Intelligence Agent to apply discriminatory criteria in the promotion, publication or brokerage of properties, this prohibition being additional and complementary to the conduct obligations provided in the Terms and Conditions of Service.
8.5 Bias reporting mechanism
The User may report to the Data Controller, through the channels described in Chapter 14, any result generated by an Artificial Intelligence component that they identify as potentially discriminatory, the Data Controller committing to investigate such report pursuant to its internal governance processes.
8.6 Consequences of non-compliance
A User's use of a Platform Artificial Intelligence component for discriminatory purposes shall constitute a material breach of this Policy and of the Terms and Conditions of Service, entitling the Data Controller to suspend or terminate the responsible User's Account, without prejudice to the liability that the latter may have under applicable non-discrimination legislation.
Chapter 9. Security, Robustness and Prevention of Misuse
9.1 Legal objective
To develop the security and robustness principle indicated in section 3.5, establishing the measures the Data Controller adopts to prevent misuse of its Artificial Intelligence components.
9.2 Controls against instruction manipulation
The Data Controller implements reasonable technical controls aimed at preventing a User or third party from manipulating, through prompt-engineering techniques or other analogous means, the behavior of an Artificial Intelligence Agent so that it acts outside the parameters defined by the Data Controller, without such controls constituting an absolute guarantee against manipulation techniques unidentified at the time of implementation.
9.3 Prevention of confidential information extraction
The Data Controller implements measures aimed at preventing a User or third party from extracting, through interaction with an Artificial Intelligence Agent, confidential information from the Platform, including system prompts, automation architectures, or other Users' information, consistent with the provisions of Chapter 14 of Title V of the Terms and Conditions of Service.
9.4 Prevention of harmful content generation
The Data Controller configures its Artificial Intelligence Agents to reject requests aimed at generating unlawful, defamatory, fraudulent content, or content contrary to the Platform's Acceptable Use Policy.
9.5 Resource usage limits
The Data Controller may implement reasonable technical limits on the volume of interactions a User can sustain with the Artificial Intelligence Agents over a determined period, for the purpose of preserving the availability and security of the Platform's infrastructure for all Users.
9.6 Collaboration with third-party Artificial Intelligence Providers
When the underlying infrastructure of an Artificial Intelligence Agent is provided by a third-party Artificial Intelligence Provider, the Data Controller additionally benefits from that provider's own security controls, without this substituting the Data Controller's own controls implemented under this Chapter.
9.7 Consequences of non-compliance
A User's attempt at manipulation, extraction of confidential information, or use of an Artificial Intelligence Agent for unlawful purposes shall constitute a material breach of this Policy, entitling the Data Controller to immediately suspend or terminate the responsible User's Account, pursuant to Title X of the Terms and Conditions of Service.
TITLE IV. SPECIFIC DISCLAIMER ON THE USE OF ARTIFICIAL INTELLIGENCE
Chapter 10. Probabilistic Nature and Technical Limitations
10.1 Legal objective
To establish, with the maximum level of clarity and specificity, the probabilistic technical nature of the Platform's generative Artificial Intelligence components, as the foundation of the disclaimer regime developed in this Title.
10.2 Technical acknowledgment of the probabilistic nature
The User expressly acknowledges and accepts that the Platform's Artificial Intelligence Agents operate through probabilistic language models, which generate responses based on statistical patterns learned from large volumes of data, and not through deterministic reasoning or a genuine understanding of the generated content.
10.3 Possibility of hallucinations
The User acknowledges that, as a consequence of the nature indicated in the preceding section, the Artificial Intelligence Agents can, in exceptional but inherent cases of this technology, generate Hallucinations, that is, incorrect, outdated information or information unsupported by real data, presented with an appearance of certainty, a phenomenon that constitutes a recognized and documented technical limitation of this category of technology worldwide, and not a defect attributable to a lack of specific diligence by the Data Controller.
10.4 Distinction regarding the Calculation Engine
Pursuant to section 1.6, the limitation described in this Chapter is proper and exclusive to the generative or conversational Artificial Intelligence Agents; the Calculation Engine, operating through fixed deterministic formulas, is not subject to the Hallucination phenomenon, without prejudice to its results depending on the accuracy of the input data pursuant to section 11.4.
10.5 Evolution of the technology
The User acknowledges that generative Artificial Intelligence technology is in constant evolution, and that the technical limitations described in this Chapter, while subject to continuous mitigation by the industry and the Data Controller itself, cannot be considered absolutely eliminated in the current state of technological development.
10.6 Consequences of non-compliance
This Chapter is declaratory and informative in nature, and the User's acknowledgment of it is a condition for using the Platform's Artificial Intelligence components; the corresponding liability regime is developed in the subsequent Chapters of this Title.
Chapter 11. Disclaimer Regarding Results and Indicators
11.1 Informational and non-binding character
Every result, response, analysis, comparison, or recommendation generated by an Artificial Intelligence Agent or by the Platform's Calculation Engine is exclusively informational in character and supportive of the User's professional decision-making, and in no case constitutes financial, tax, legal, accounting or personalized investment advice, nor a guarantee, promise or certain projection of future result, performance or capital appreciation.
11.2 Exclusion of liability for Hallucinations
The Data Controller shall not be liable for damages deriving from a decision adopted by the User or an end client based on a Hallucination generated by an Artificial Intelligence Agent, provided the Data Controller has reasonably complied with the governance, human oversight and transparency mechanisms described in Titles II and III of this Policy.
11.3 User's duty of verification
The User undertakes to exercise their own professional judgment when using the results generated by the Artificial Intelligence Agents, and to verify, when the nature of the decision involved reasonably warrants it (particularly in real-estate transactions of significant value), the relevant information through additional sources, before conveying such information to their client or making a decision based on it.
11.4 Accuracy of Market Indicators dependent on source data
The User acknowledges that the accuracy of the Market Indicators generated by the Calculation Engine depends on the accuracy, currency and completeness of the market data underlying them, which comes from public sources, the Platform's own database, and the information loaded by the Users themselves, without the Data Controller being able to absolutely guarantee the accuracy of such external sources.
11.5 Applicable quantitative limit
The disclaimer established in this Chapter is understood without prejudice to, and subject to, the quantitative limit of liability established in section 39.2 of the Terms and Conditions of Service, which applies to every claim related to the use of the Platform's Artificial Intelligence components, save the exceptions provided there.
11.6 Consequences of non-compliance
This Chapter constitutes an essential and non-negotiable condition for accessing and using the Platform's Artificial Intelligence components, the User acknowledging that the prices of the Plans including those components have been determined considering this disclaimer regime.
Chapter 12. Absence of Automated Decisions with Definitive Legal Effects
12.1 General declaration
The Data Controller declares that no Artificial Intelligence component of the Platform is designed to, nor has the effect of, adopting an Automated Decision with Definitive Legal Effects regarding an individual, without human intervention.
12.2 Supportive nature, not substitution
The results generated by the Artificial Intelligence Agents and the Calculation Engine constitute supporting inputs for the Real-Estate Agent, in the exercise of their own professional judgment, to adopt or recommend a decision to their client; in no case do such results substitute the professional human intervention the corresponding real-estate transaction requires.
12.3 Illustrative examples
By way of illustration, the following do not constitute Automated Decisions with Definitive Legal Effects: the generation of a return-on-investment calculation; the suggestion of properties by the recommendation engine; an Artificial Intelligence Agent's response to an end client's inquiry; and the generation of a Market Indicator regarding a determined zone. In all cases, the final decision to buy, sell, lease, invest or execute any legal act belongs exclusively and freely to the individual involved, with the advice of the corresponding Real-Estate Agent.
12.4 Consequences of non-compliance
This declaration is of general application; in the exceptional event that, due to the Platform's evolution, a feature that does constitute an Automated Decision with Definitive Legal Effects were incorporated, the Data Controller will update this Policy and the Full Privacy Notice pursuant to their respective amendment mechanisms, prior to the implementation of such feature.
Chapter 13. Professional Responsibility of the Real-Estate Agent Toward their Client
13.1 Exclusive advisory relationship between the Real-Estate Agent and their client
The Data Controller maintains no real-estate, financial or other professional advisory relationship with the Real-Estate Agents' end clients; that advisory relationship is established exclusively between the Real-Estate Agent and their client, the Platform being a technology tool supporting that relationship.
13.2 Obligation of professional contextualization
The Real-Estate Agent who uses results generated by the Platform's Artificial Intelligence components in their interaction with a client is solely responsible for contextualizing such information according to their experience and professional judgment, for communicating to their client the informational and non-binding character of such results, and for complementing it with their own expert judgment before conveying it as a recommendation.
13.3 Independence of professional liability
The professional, ethical and, where applicable, legal liability the Real-Estate Agent has toward their client, under real-estate industry legislation and practice, is independent of the contractual relationship between that Agent and the Data Controller, and is not altered, substituted or diminished by the use of the Platform's Artificial Intelligence components.
13.4 Consequences of non-compliance
A Real-Estate Agent's non-compliance with the professional contextualization obligations indicated in this Chapter is a matter foreign to the contractual relationship between that Agent and the Data Controller, without prejudice to the Data Controller becoming aware of such non-compliance through the reporting mechanism described in Chapter 14, and to the contractual consequences that, where applicable, apply under the Terms and Conditions of Service.
TITLE V. INCIDENT MANAGEMENT AND CONTINUOUS IMPROVEMENT
Chapter 14. Reporting and Management of Artificial Intelligence Incidents
14.1 Legal objective
To establish the procedure through which the User can report an Artificial Intelligence Incident, and the Data Controller's corresponding handling commitment, developing the accountability principle indicated in section 3.6.
14.2 Reporting channel
The User may report an Artificial Intelligence Incident (a manifestly erroneous, discriminatory, unsafe result, or one that substantially departs from expected behavior) via the email address legal@ancuria.com, or through the specific reporting mechanism that, where applicable, is enabled within the Platform's own interface next to the responses generated by the Artificial Intelligence Agents.
14.3 Information to provide in the report
The User should provide, to the extent possible, a description of the Incident, the approximate date and time it occurred, and, where technically possible, the specific interaction that gave rise to the reported result, in order to facilitate its investigation by the Data Controller.
14.4 Handling procedure
The Data Controller will review every report received under this Chapter, classifying it by severity and determining the corresponding corrective actions, which may include, by way of illustration: adjusting the instructions of the Artificial Intelligence Agent involved; incorporating the case into the continuous improvement processes indicated in Chapter 15; or, in cases of significant severity, temporarily suspending the involved component while the root cause is investigated and corrected.
14.5 Handling period
The Data Controller will endeavor to give an initial response to the User submitting a report under this Chapter within a reasonable period not exceeding [POR DEFINIR] business days, without that period constituting a guarantee of definitive resolution of the Incident within it, given the technical nature of the investigation that may be necessary.
14.6 Escalation of severe incidents
Every Artificial Intelligence Incident involving a potentially discriminatory result, the exposure of confidential information, or a significant security risk will be escalated on a priority basis pursuant to the Data Controller's internal governance processes described in Chapter 4.
14.7 Consequences of non-compliance
Failure to reasonably attend to a well-founded report of an Artificial Intelligence Incident may be considered a relevant factor in determining the Data Controller's diligence in a claim related to that Incident.
Chapter 15. Continuous Improvement and Model Updates
15.1 Legal objective
To develop the continuous improvement principle indicated in section 3.8, describing the Data Controller's approach to the evolution of its Artificial Intelligence components based on the Platform's operational experience.
15.2 Incorporation of feedback
The Data Controller incorporates, according to its internal processes, the Feedback provided by Users under Chapter 17 of the Terms and Conditions of Service, as well as the findings deriving from the Incident reports indicated in Chapter 14, into its processes of adjustment and improvement of the Platform's Artificial Intelligence components.
15.3 Periodic performance review
The Data Controller conducts periodic reviews of the performance of its Artificial Intelligence components, evaluating indicators such as the frequency of Incident reports, overall User satisfaction, and the evolution of the technical capabilities available in the industry.
15.4 Adoption of industry best practices
The Data Controller undertakes to stay informed about the evolution of international best practices in responsible Artificial Intelligence, incorporating them into its operation as they are applicable and proportional to the Platform's size and nature.
15.5 Consequences of non-compliance
The provisions of this Chapter describe the Data Controller's continuous improvement commitment for transparency purposes toward Users, investors and auditors, without this generating specific result obligations other than those expressly agreed in the Terms and Conditions of Service.
Chapter 16. Third-party Artificial Intelligence Providers
16.1 Legal objective
To inform about the use of technology infrastructure of third-party Artificial Intelligence Providers in the operation of the Platform's Artificial Intelligence components, and to specify the corresponding liability regime.
16.2 Use of third-party infrastructure
The Data Controller may base the operation of its Artificial Intelligence Agents, wholly or partially, on language models and infrastructure provided by third-party Artificial Intelligence Providers, whose identity can be consulted through the website www.ancuria.com or the email address legal@ancuria.com.
16.3 Nature of the relationship with such providers
Third-party Artificial Intelligence Providers act, regarding the processing of Personal Data that such infrastructure implies, as data processors pursuant to Chapter 10 of the Full Privacy Notice; regarding the intellectual property of their base models, they retain the ownership recognized in section 14.3 of the Terms and Conditions of Service.
16.4 Data Controller's liability toward the User
Independently of whether the underlying infrastructure of an Artificial Intelligence Agent is provided by a third-party Artificial Intelligence Provider, the Data Controller is the sole point of contact and liability toward the User regarding that Agent's operation within the Platform, without prejudice to its right of recourse against that provider under the corresponding commercial agreements, when the cause of an Incident is attributable to the latter.
16.5 Continuity upon provider change
The Data Controller may substitute one third-party Artificial Intelligence Provider for another, when it considers it convenient for the Platform's operation, quality or security, without such substitution, by itself, requiring a material amendment of this Policy, unless it materially alters the capabilities or limitations informed to the User.
16.6 Consequences of non-compliance
Failure to inform about a relevant change of third-party Artificial Intelligence Provider that materially affects the Platform's capabilities may be considered a breach of the transparency principle indicated in section 3.2.
TITLE VI. FINAL PROVISIONS
Chapter 17. Relationship with Other Documents of Ancuria's Legal Framework
17.1 Documentary hierarchy on Artificial Intelligence matters
In everything specifically relating to the use of Artificial Intelligence within the Platform, this Policy prevails over the Terms and Conditions of Service and over the Full Privacy Notice; in everything not provided for in this Policy, those documents apply in a supplementary manner under their own terms.
17.2 Main cross-references
Without prejudice to the specific references contained throughout this Policy, the following stand out as main cross-references: Chapter 14 (Intellectual Property over Artificial Intelligence) and Chapter 40 (Specific Disclaimer on Artificial Intelligence Results and Market Indicators) of the Terms and Conditions of Service; and Chapter 9 (Specific Purposes of Processing in Artificial Intelligence Processes) of the Full Privacy Notice.
17.3 Consequences of non-compliance
Any interpretation of any of the documents of Ancuria's legal framework in contravention of the hierarchy established in this Chapter shall be deemed invalid, the interpretation pursuant to what is provided here prevailing.
Chapter 18. Validity and Amendments
18.1 Entry into force
This Policy shall enter into force on the date stated on the cover page of this document.
18.2 Power to amend
The Data Controller may modify this Policy at any time, particularly to reflect the incorporation of new Artificial Intelligence components, changes of third-party Artificial Intelligence Providers, or the evolution of industry best practices.
18.3 Notification mechanism
Every material amendment to this Policy will be notified to the User pursuant to the same mechanism provided in Chapter 4 of the Terms and Conditions of Service.
18.4 Consequences of non-compliance
Failure to notify a material amendment under this Chapter shall result in that amendment not being enforceable against the affected User until its effective notification.
Chapter 19. Final Provisions
19.1 Severability
Should any provision of this Policy be declared null, illegal or unenforceable by a competent authority, such declaration shall not affect the validity and enforceability of the remaining provisions.
19.2 Survival
The provisions of Title IV (Specific Disclaimer on the Use of Artificial Intelligence) shall survive termination of the contractual relationship between the User and the Data Controller, with respect to any claim related to events occurring during the validity of that relationship.
19.3 Governing language
In the event of translation into another language, the Spanish (Mexico) version shall prevail for interpretation purposes in case of discrepancy.
19.4 Interpretation according to guiding principles
In case of any ambiguity in the interpretation of this Policy, the interpretation most consistent with the guiding principles established in Chapter 3 must be favored.
Version Control Annex
| Version | Effective date | Description of the change |
|---|---|---|
| 1.0 | September 22, 2026 | Initial version of the document |