# Responsible Artificial Intelligence Use Policy > Principles, responsibilities, and restrictions governing Growth Station AI-powered features. - Canonical URL: https://docs.growthstation.app/docs/legal/en/responsible-ai-use - Language: en - Last updated: 2026-08-17T12:49:19.123Z - Product: Plataforma Growth Station - Maintained by: Privacy — Growth Machine - Editorial review: 2026-04-14 Growth Station — Responsible AI Use Policy [#growth-station--responsible-ai-use-policy] **Last updated:** April 14, 2026\ **Document owner:** Petro Liporace Cardoso de Souza\ **Classification:** Public document > **Translation notice:** This English version is provided for convenience. If it conflicts with the original Portuguese document, the Portuguese version prevails. *** 1. Purpose [#1-purpose] This Policy establishes guidelines for the ethical, safe, and responsible use of Artificial Intelligence (AI) features available on the Growth Station Platform, with transparency, human oversight, and compliance with Brazilian law. *** 2. Scope [#2-scope] This Policy applies to every AI-powered Platform feature, including: | Feature | Description | | ---------------------------------- | ----------------------------------------------------------------------------- | | **AI prospecting agent** | Automated contact and follow-up with Leads through email, WhatsApp, and calls | | **VoIP call analysis** | Automated call quality and performance assessment | | **Call transcription** | Conversion of call audio into text | | **WhatsApp audio transcription** | Conversion of audio messages into text | | **Media description** | Analysis and description of images and vCards received through WhatsApp | | **Message generation** | Drafting emails and WhatsApp messages | | **Agent self-improvement** | Automated behavior adjustments based on outcomes | | **Lead enrichment** | Automated addition of complementary lead data | | **Course and training assistance** | AI-powered suggestions and support for User training activities | *** 3. Responsible AI Principles [#3-responsible-ai-principles] 3.1 Transparency [#31-transparency] * The Project (customer company) must inform its Leads when communications are generated or sent by AI. * A complete history of AI agent decisions and actions is retained and available to the Project. * Interactions with AI models are traced through Langfuse for audit purposes. 3.2 Human Oversight [#32-human-oversight] * **The AI agent does not replace human judgment.** AI-generated analyses, suggestions, and classifications are recommendations. * The Project can configure **scope rules** that restrict the agent's actions. * Authorized Users can review, approve, edit, or reject actions suggested by the agent. * The Project can **pause or disable** the AI agent at any time. 3.3 Non-Discrimination [#33-non-discrimination] * AI features must not be used to generate discriminatory content based on race, gender, sexual orientation, religion, disability, origin, age, or any other characteristic protected by law. * The language models used by the Platform (OpenAI and Groq) include built-in content filters. 3.4 Accuracy and Limitations [#34-accuracy-and-limitations] * Growth Machine **does not guarantee** that AI-generated content will be accurate, complete, or suitable. * Language models may generate inaccurate information or "hallucinations" (confabulations). * The User remains **ultimately responsible** for all content sent to Leads, including AI-generated content. * Call and audio transcripts may contain inaccuracies, particularly in noisy environments. 3.5 Privacy and Data Protection [#35-privacy-and-data-protection] * Data processed by AI is subject to the Platform's Privacy Policy and DPA. * Data is sent to AI providers (OpenAI and Groq) through their APIs under DPAs that prohibit its use for model training. * Only data necessary for the specific feature is sent to the AI model. *** 4. Permitted Use [#4-permitted-use] AI features may be used for: 1. **Legitimate sales prospecting** — Contacting qualified leads within the Project's business scope. 2. **Qualification and prioritization** — Analyzing lead data to suggest next actions. 3. **Message generation** — Drafting emails and personalized prospecting messages. 4. **Performance analysis** — Assessing calls and interactions to improve processes. 5. **Repetitive task automation** — Follow-ups, scheduling, and status updates. *** 5. Prohibited Use [#5-prohibited-use] AI features must not be used for: 1. **Spam or unsolicited bulk communications** — Automated delivery without a legal basis or relevance to the recipient. 2. **Deception or fraud** — Leading a person to believe they are communicating with a human when they are interacting with an AI agent, where disclosure is legally required. 3. **Illegal content** — Generating threats, harassment, defamation, phishing, or any unlawful content. 4. **Discrimination** — Selecting, excluding, or prioritizing leads based on protected characteristics. 5. **Harmful automated decisions** — Making decisions that significantly affect data subject rights without human review. 6. **Model reverse engineering** — Attempting to extract, copy, or replicate underlying AI models. 7. **Sensitive data processing** — Deliberately sending sensitive personal data, as defined in Article 11 of the LGPD, to an AI model without an appropriate legal basis. 8. **Manipulation or coercion** — Using AI to manipulate or coerce Leads into making decisions. *** 6. AI Agent Configuration and Control [#6-ai-agent-configuration-and-control] 6.1 Project Responsibilities [#61-project-responsibilities] The Project (customer company) is responsible for: | Responsibility | Description | | ---------------------------- | --------------------------------------------------------------------------- | | **Defining scope rules** | Configuring operating limits, tone of voice, language, and restrictions | | **Monitoring agent actions** | Periodically reviewing event and decision history | | **Managing credits** | Maintaining sufficient credits for agent operation | | **Ensuring a legal basis** | Ensuring AI use in communications with Leads has an appropriate legal basis | | **Overseeing content** | Reviewing and approving communication templates and rules | 6.2 Technical Controls [#62-technical-controls] | Control | Description | | ------------------------------- | ------------------------------------------------------------------------------- | | **Credit limits** | The agent is automatically paused when credits are insufficient | | **Scope rules** | Configurable operating limits and communication tones | | **Audit history** | Complete record of agent decisions and actions | | **Tracing (Langfuse)** | Traceability of calls to AI models | | **Supervised self-improvement** | Automated adjustments based on metrics and subject to configurable limits | | **Automatic freeze** | Scheduled routines that identify and pause agents with inadequate configuration | *** 7. Data Subject Rights Related to AI [#7-data-subject-rights-related-to-ai] Under Article 20 of the LGPD: | Right | Implementation | | ------------------------------------------- | ---------------------------------------------------------------- | | **Request review of an automated decision** | A Lead or Project may request human review of an AI agent action | | **Obtain information about criteria** | The Project can access the agent's decision history | | **Challenge a decision** | The Project can reverse an agent action | *** 8. AI Providers [#8-ai-providers] | Provider | Use | Data policy | | ------------ | ------------------------------- | ----------------------------------------------------------------------- | | **OpenAI** | Primary language model provider | Data sent through the API is not used for training; DPA in place; SOC 2 | | **Groq** | Fallback provider | Same scope of use; DPA in place | | **Langfuse** | AI interaction monitoring | Tracing data, including prompts, responses, and metrics; DPA in place | *** 9. AI Incidents [#9-ai-incidents] The following events must be reported as incidents: 1. The AI agent sends inappropriate, offensive, or false content to Leads. 2. The AI agent exceeds its configured scope. 3. Data from one Project is exposed to another Project through AI. 4. The AI generates discriminatory or illegal content. 5. Prompt injection or external manipulation of the agent occurs. **Process:** Report the event under the Incident Response Plan → classify it with at least Medium severity → begin a priority investigation. *** 10. Enforcement [#10-enforcement] Use of AI in breach of this Policy may result in: 1. Disabling the Project's AI agent. 2. Suspending AI features. 3. Suspending or terminating the account under the Terms of Use. 4. Civil liability under applicable law. *** *This document was prepared based on Brazil's LGPD (Law No. 13,709/2018), the Brazilian Civil Rights Framework for the Internet (Law No. 12,965/2014), and the OECD principles on responsible AI.*