These AI Supplemental Terms ("AI Terms") supplement the GenZ OS Terms of Service and apply when Customer accesses an AI-assisted or agentic feature. If these AI Terms conflict with the Terms of Service concerning an AI feature, these AI Terms control for that subject.
AI Features
AI Features may interpret natural-language prompts, project files, images, code, requirements, repository context, or other Customer Content and may generate or modify architecture, code, APIs, schemas, documentation, tests, wireframes, prototypes, tasks, or recommendations. Agents may plan and execute multiple generation steps, call configured tools, test code, or iterate within permissions established by Customer and Provider.
Model Choice and Third-Party Providers
The Service may offer different models or model families, including commercial and open or hosted models. Availability, context limits, price, latency, data location, retention, and output behavior may differ. Customer authorizes transmission of the information reasonably necessary to the model or hosting provider selected by Customer or configured for the feature.
A model or integration provider may impose separate acceptable-use terms and may suspend or change access. Provider may substitute routing where needed for availability, safety, law, or contractual requirements, subject to an Order Form and any disclosed data-location restriction.
Inputs, Rights, and Restrictions
Customer is responsible for Inputs and represents that it has the rights and lawful authority necessary to process them. Customer must not submit personal data, confidential source code, credentials, trade secrets, export-controlled data, or third-party material unless the use is authorized and the selected model, deployment, and contract are appropriate.
Customer must not attempt prompt injection against Provider or third-party systems; extract hidden prompts, weights, training data, or non-public controls; defeat safety systems; or use AI Features for conduct prohibited by the Acceptable Use Policy.
Output Rights and Non-Exclusivity
Rights in Output are addressed in the Terms of Service. Output may not qualify for intellectual-property protection and may be the same as or similar to output provided to another user. Provider does not guarantee exclusivity, originality, non-infringement, patentability, copyrightability, or freedom to operate.
Open-source software, third-party APIs, packages, datasets, standards, and code fragments may remain governed by their own licenses and notices. Customer is responsible for software-composition analysis and compliance before copying, distributing, licensing, or deploying Output.
Accuracy, Security, and Human Review
AI systems may hallucinate, omit requirements, introduce vulnerabilities, select outdated packages, misapply a license, generate insecure defaults, or produce code that fails under real conditions. Automated tests may be incomplete and passing tests do not establish security, legal compliance, accessibility, performance, or suitability.
Before use or deployment, Customer must apply review proportionate to risk, including:
validation against approved requirements and architecture;
qualified code review and secure-development review;
dependency, provenance, license, and intellectual-property analysis;
secret, malware, vulnerability, and software-supply-chain scanning;
unit, integration, system, regression, performance, resilience, and accessibility testing;
privacy, security, data-protection, and threat assessments;
sandboxing and staged deployment with rollback and monitoring; and
competent human approval for regulated, safety-critical, or consequential use.
Agentic Actions and Tool Permissions
Customer is responsible for tools, repositories, APIs, environments, credentials, permissions, and approval gates available to an agent. Customer should apply least privilege, separate development from production, require human approval for external or irreversible actions, preserve logs, and maintain backups and rollback mechanisms.
An agent's plan, tool call, test, or apparent success is not a representation that an action is authorized, safe, complete, or correct. Customer must review proposed changes before merging, releasing, deleting data, altering infrastructure, or making a production change.
Training and Improvement
Unless an Order Form or clear in-product notice expressly states otherwise, Provider does not use Customer Content to train a third-party general-purpose model. Provider may process feedback and de-identified or aggregated operational data to evaluate, secure, and improve AI Features. Any optional program that uses identifiable Customer Content for model improvement must be presented with appropriate notice and choice.
Transparency and AI Literacy
Customer must provide legally required notice that an application, interaction, or material is AI-generated or AI-assisted. Customer must ensure that personnel operating or supervising AI systems receive instruction appropriate to their role, technical knowledge, context, and foreseeable impact. Customer must preserve records sufficient to explain material workflows where required.
Regulated, High-Risk, and Prohibited Uses
Customer is responsible for determining whether it is a provider, deployer, developer, distributor, importer, operator, or other regulated actor under applicable AI law. GenZos.ai's role as a software-development platform does not transfer Customer's obligations for an application Customer places on the market or uses.
Customer must not use AI Features for prohibited AI practices or as the sole basis for a legally significant decision about a person. A high-risk or regulated use requires written approval where the standard Service is not expressly designed for it, along with risk management, data governance, validation, documentation, logging, human oversight, transparency, monitoring, incident response, and conformity or impact assessment where applicable.
Privacy, Evaluations, and Benchmarking
Prompts, repository context, files, and Output may contain personal or confidential information. Customer must minimize data, configure retention and access, select an appropriate model and deployment, and execute the Data Processing Addendum where Provider processes Customer Personal Data. Customer should use synthetic or de-identified data for testing where practical.
Public benchmarking must describe the tested model, version, configuration, prompt set, sample size, date, and material limitations and must not misrepresent results or disclose confidential information. Provider may restrict tests that threaten security or materially disrupt the Service.
Suspension, Changes, and Disclaimer
Provider may restrict an AI Feature, model, tool, or Output where needed to address unlawful use, safety, security, third-party requirements, intellectual-property risk, or material platform harm. Model availability and behavior may change. Provider will provide notice of a material paid-feature change where reasonably practicable.
AI FEATURES AND OUTPUT ARE SUBJECT TO THE DISCLAIMERS AND LIABILITY LIMITS IN THE TERMS OF SERVICE. CUSTOMER ACKNOWLEDGES THAT AI OUTPUT REQUIRES INDEPENDENT REVIEW AND THAT CUSTOMER, NOT PROVIDER, CONTROLS THE DECISION TO USE, MERGE, DISTRIBUTE, OR DEPLOY OUTPUT.