ARTIFICIAL INTELLIGENCE POLICY — LEXBUDDY
Version 1.1 | Last updated: June 4, 2026
This Artificial Intelligence Policy (hereinafter, the “AI Policy”) describes LexBuddy’s approach to the development and deployment of artificial intelligence technologies on its platform. Its purpose is to provide transparency about how LexBuddy operates, what principles guide its development, and what commitments it assumes toward its users.
This AI Policy should be read together with the LexBuddy Terms of Use and Privacy Notice.
1. NATURE OF THE PLATFORM
LexBuddy is a platform that supports legal work using artificial intelligence — specifically large language models (Large Language Models or LLMs) developed by third parties — to process queries, generate documents, and assist the user in legal tasks.
LexBuddy is a SUPPORT TOOL. The platform does not replace the lawyer’s professional judgment or legal advice. Generated outputs are preliminary and must be reviewed by the user before use.
2. GUIDING PRINCIPLES
The development and operation of LexBuddy are guided by the following principles:
2.1 Support, not substitution. Artificial intelligence is a tool that enhances the user’s work; it does not replace their judgment, professional discretion, or responsibility.
2.2 Human oversight. LexBuddy does not make legal decisions on its own. Every final decision belongs to the user, who must review, validate, and, where applicable, correct outputs before using them.
2.3 Transparency. LexBuddy seeks to be clear about the origin and limitations of its responses, distinguishing when they are based on verifiable sources and when they derive from the model’s general knowledge.
2.4 Confidentiality. Information the user processes through the platform is treated with high standards of confidentiality and is not used for purposes other than providing the service.
2.5 Respect for professional responsibility. LexBuddy recognizes that its users — especially lawyers — are subject to their own professional duties, including professional secrecy. The platform is designed to be consistent with such duties.
3. HOW LEXBUDDY WORKS
3.1 Underlying models. LexBuddy uses language models developed by external providers specialized in artificial intelligence. The platform may modify, replace, or supplement these models over time to improve the service.
3.2 Internal legal repository. LexBuddy maintains a structured repository of laws and other legal sources, which it uses as a reference to support legal responses. LexBuddy seeks to keep this repository updated in accordance with applicable legislation.
3.3 Search in official sources. For jurisprudential criteria and other queries requiring real-time information, LexBuddy consults official sources directly (Supreme Court of Justice of the Nation, Official Gazette of the Federation, among others).
3.4 Source traceability. LexBuddy seeks to include references to the sources that support its responses, facilitating independent verification of the information by the user.
4. KNOWN LIMITATIONS
LexBuddy explicitly acknowledges the following limitations inherent in the artificial intelligence technologies it uses:
4.1 Possible errors. Language models may generate inaccurate, incomplete, or incorrect responses. They may incur what is known as “hallucinations”: statements that sound plausible but do not correspond to reality.
4.2 Outdated information. Language models’ knowledge has a cutoff date. Recent information on legal reforms, jurisprudential criteria, or authority decisions may not be reflected accurately.
4.3 Non-infallible heuristics. LexBuddy applies criteria to determine when a query requires formal legal grounding, search in official sources, or an orientational response from general knowledge. These criteria may not be infallible.
4.4 Potential biases. Language models may reflect biases present in the data on which they were trained. LexBuddy seeks to mitigate these biases but cannot eliminate them entirely.
4.5 Verification recommended. For any professional, contractual, or procedural use of LexBuddy outputs, independent verification by the user is recommended.
5. DATA AND NO TRAINING
5.1 No training. LexBuddy does NOT use user information — conversations, documents, data, or any other — to train, improve, or develop artificial intelligence models.
5.2 Extended commitment to providers. This restriction also applies to external artificial intelligence providers used by the platform. LexBuddy contractually requires its providers not to retain or use processed inputs and outputs for training purposes.
5.3 No human access by providers. External artificial intelligence providers process user data automatically. The platform contractually requires that such providers do not allow human review of processed information, unless a court in the United States of America under the CLOUD Act (2018) or Mexico so requires. This restriction applies to the artificial intelligence model provider. Cloud infrastructure providers used by LexBuddy operate in accordance with their own security policies and applicable certifications.
5.4 Restricted internal access. Access by the LexBuddy team to user information is restricted to strictly necessary cases — such as resolution of security incidents or authorized technical support — and is recorded in audit logs.
5.5 User information for service improvement. LexBuddy may analyze aggregated and anonymized platform usage data (such as response times, frequency of features used, technical errors) to improve the service. This analysis never includes specific content from the user’s conversations or documents.
6. EPISTEMIC TRANSPARENCY
LexBuddy seeks to distinguish, to the extent possible, the origin of its responses:
6.1 Responses with verifiable grounding. When a response is based on the internal legal repository or searches in official sources, LexBuddy includes references or links to such sources to facilitate verification by the user.
6.2 Orientational responses. When a response comes from the model’s general knowledge and is not based on a specific source, LexBuddy seeks to indicate this, recommending that the user verify against official sources before professional use.
6.3 Limitations of traceability. The platform’s ability to distinguish the origin of each fragment of a response is a feature under development. LexBuddy works continuously to improve this capability.
7. HUMAN OVERSIGHT AND RESPONSIBILITY
7.1 User decisions. LexBuddy does not make legal decisions autonomously. Every platform output is preliminary and requires validation by the user.
7.2 Actions on external tools. When LexBuddy offers action capabilities on external tools (for example, preparing an email in an inbox, identifying files in cloud storage), the platform proposes — the user reviews and executes. No irreversible action is taken without the user’s express confirmation.
7.3 Professional responsibility. Professional responsibility for documents, briefs, contracts, or consultations the user produces with LexBuddy’s support belongs exclusively to the user.
8. REGULATORY COMPLIANCE
8.1 Applicable legal framework. LexBuddy operates in accordance with applicable Mexican legislation, including the Federal Law on Protection of Personal Data Held by Private Parties (LFPDPPP) and other relevant provisions.
8.2 Regulatory evolution. LexBuddy recognizes that the regulatory framework applicable to artificial intelligence is evolving, both in Mexico and internationally. The platform commits to adapting its practices and policies as such regulation develops.
8.3 Reports and audits. LexBuddy commits to cooperating with competent authorities to the extent required by law, without compromising the confidentiality of user information beyond what is legally required.
8.4. Ethical use. LexBuddy trains its team in the ethical and responsible use of artificial intelligence and maintains internal policies in this regard. The platform is designed so that users are aware at all times that they are interacting with an artificial intelligence system, not a human professional.
9. MODEL UPDATES
9.1 Model changes. LexBuddy may update, replace, or incorporate new artificial intelligence models to improve the service.
9.2 Permanent commitments. Regardless of the model used, LexBuddy’s commitments regarding non-training, confidentiality, restricted access, and other obligations contained in this AI Policy remain in effect.
9.3 Notice of material changes. LexBuddy will timely notify users of material changes to the underlying model or its artificial intelligence practices.
10. RISK MANAGEMENT
10.1 Risk identification. LexBuddy continuously identifies and evaluates risks associated with the use of artificial intelligence in a legal context, including errors in outputs, biases, excessive user reliance, and confidentiality risks.
10.2 Mitigation measures. The platform implements mitigation measures, such as:
-
Persistent disclaimers reminding users of the preliminary nature of outputs.
-
Source traceability when possible.
-
Technical and organizational restrictions on internal access to data.
-
Periodic security and privacy audits.
-
Continuous development of criteria to distinguish when a query requires formal legal grounding.
10.3 Incident reporting. If an incident is identified that affects the security or confidentiality of user information, LexBuddy will notify in accordance with its Privacy Notice and applicable legislation.
11. UPDATES TO THIS POLICY
LexBuddy may update this AI Policy as its practices, technology, or applicable regulation evolve. Updates will be published on the platform and, when they constitute material changes, users will be notified with reasonable advance notice.
12. CONTACT
For any questions or comments about this AI Policy, you may contact us at: