Responsible Use Policy

1. Services Covered by This Policy

These guidelines apply whenever you access or use FildraAI products, whether through the website, mobile applications, APIs, integrations, playgrounds, or internal tools made available to you.

2. How to Use FildraAI Responsibly

• Review generated answers before using them in operational, financial, safety, or customer-facing decisions.
• Provide accurate context, such as crop, location, season, language, image quality, and relevant constraints.
• Do not rely on FildraAI as an emergency system or as the only source for high-risk decisions.
• Use the service in a way that respects other users, local communities, and applicable laws.
• Report outputs that appear unsafe, misleading, discriminatory, or inconsistent with local reality.

Service-Specific Notes

• FieldGuide and FieldState: Use guidance and farm memory as planning support. Verify important actions with local expertise, records, labels, and regulations.
• FieldVision: Image diagnosis depends on image quality, crop coverage, lighting, symptoms, and model limitations. Confirm uncertain results before treatment.
• FieldAudio: Transcription and translation may be imperfect, especially with accents, noise, dialects, mixed languages, or technical terms. Review important translations.
• fieldmap and location tools: Maps, weather context, and spatial outputs are decision-support information and may not reflect real-time conditions or legal boundaries.
• APIs: Use API keys only for authorized applications and within your plan, quota, scope, and security responsibilities.

3. Human Review and High-Risk Decisions

FildraAI is designed to support decisions, not make final decisions on your behalf. Higher-risk uses require stronger human review, local validation, and where appropriate, qualified professional advice.

4. Inputs, Uploads, and Data Hygiene

The quality and safety of FildraAI outputs depend heavily on what users submit. Use care when uploading, recording, typing, or syncing information.

• Provide accurate context: Include relevant crop, location, language, season, symptoms, image angle, and management details where useful.
• Use clear uploads: Blurry images, noisy audio, incomplete records, or wrong location context may reduce output quality.
• Avoid sensitive personal data: Do not submit government ID numbers, passwords, banking credentials, private medical information, or other unnecessary sensitive data.
• Respect permissions: Do not upload images, audio, documents, locations, or records that you do not have the right to use or share. Specifically: do not upload images you found online, scraped, or downloaded from another product unless you own the copyright or have an explicit licence; do not upload someone else’s voice recording unless they have given you clear, informed consent to record and share it; do not upload copyrighted material (book pages, news photos, music, paid datasets) belonging to a third party. Accounts that repeatedly upload infringing or non-consensual media may be suspended or terminated.
• Be careful with people in media: Avoid uploading identifiable people, private conversations, children, or bystanders unless you have appropriate permission.
• Voice recordings need consent: When you use FieldAudio to record speech, you are responsible for confirming that everyone whose voice is captured has knowingly agreed to be recorded for this purpose. Some jurisdictions require all-parties consent; others only one party. Know which applies to where you are. We never use your recordings to train voice-cloning or speaker-recognition models, but local recording-consent law applies to you regardless of what we do with the audio.
• Image copyright matters: FieldVision is designed for photographs you take of your own fields, crops, livestock, and equipment. If you submit images you did not take, you must have a written licence or other clear permission to use them in this product. Submitting someone else’s copyrighted image can expose you to a takedown notice, a copyright claim, or account suspension.

5. Fair Use, API Access, and Platform Integrity

We provide FildraAI services to support legitimate product, research, agricultural, translation, and operational workflows. To maintain service quality and protect our systems, users must access the platform responsibly.

5.1 Acceptable Use

• Using FildraAI tools for normal account, app, website, research, translation, diagnosis, mapping, or workflow support.
• Using APIs within your authorized keys, scopes, quotas, and plan limits.
• Uploading reasonable files, images, audio, prompts, and records that you have the right to submit.
• Sharing outputs within your organization or project where permitted by your plan and applicable law.

5.2 Prohibited Activities

• Unauthorized Automated Access: Using bots, scrapers, crawlers, scripts, or automation to access services without written permission.
• Data Harvesting: Systematically extracting content, outputs, model behavior, knowledge base data, translations, media, or platform information.
• Excessive or Abusive Requests: Sending requests that exceed plan limits, degrade reliability, bypass quotas, or resemble denial-of-service activity.
• Reverse Engineering: Attempting to extract, replicate, probe, or reverse-engineer models, prompts, training data, infrastructure, security controls, or algorithms.
• Unauthorized Resale or Redistribution: Reselling, sublicensing, republishing, or redistributing FildraAI services, outputs, API access, or content without permission.
• Credential and Access Abuse: Sharing login credentials, leaking API keys, bypassing access controls, impersonating others, or using another user's account.
• Illegal, Harmful, or Rights-Violating Use: Using services to violate laws, infringe rights, deceive others, harass people, or process content you are not authorized to use.
• Security Testing Without Permission: Scanning, penetration testing, vulnerability probing, load testing, or stress testing FildraAI systems without written authorization.
• Third-Party Personal Data Without Consent: Submitting personal data of other people (including photographs of identifiable persons) without their informed consent.
• Prompt Injection or Jailbreak Attempts: Crafting inputs designed to override model instructions, bypass safety filters, extract system prompts, or coerce the AI into producing prohibited content.
• Regulatory or Biosecurity Violations: Using FildraAI to generate or apply content that violates South African or your local agricultural, biosecurity, pesticide, livestock, or food-safety regulations.

5.3 Rate Limits & Usage Monitoring

We may use rate limits, quotas, plan scopes, logging, and automated abuse detection to protect service quality:

• Requests per minute, hour, day, or billing cycle based on plan and endpoint type.
• File size, upload frequency, audio length, media processing, and storage limits.
• API scopes that limit which services a key may call.
• Automated detection of abnormal usage, scraping, key leakage, or bot-like behavior.

6. Content Ownership and Output Use

6.1 Your Inputs and Data

You retain ownership of information, images, audio, text, records, and other content you submit, subject to any rights held by others. By using our services, you grant FildraAI the permissions needed to:

• Process your inputs to provide services, outputs, storage, syncing, translation, analysis, and support.
• Operate, secure, debug, and improve our services as described in our Privacy Policy and Terms.
• Use aggregated, de-identified operational metadata (request counts, latency, error rates, feature usage) for reliability and non-AI service improvement. We do not use the content you upload — photos, voice recordings, prompts, FieldState records — to train, fine-tune, or build our AI models. Our crop-diagnosis models are trained on publicly available agricultural datasets and on data we have separately licensed for that purpose.

6.2 FildraAI Services, Content, and Technology

FildraAI services, interfaces, models, prompts, knowledge bases, workflows, source code, documentation, branding, and methodologies are owned by FildraAI or its licensors. You may not:

• Copy, modify, resell, sublicense, or create derivative works from our technology or service content without permission.
• Extract or harvest knowledge bases, model behavior, prompts, training data, or proprietary workflows.
• Use FildraAI outputs or access patterns to train, benchmark, or build competing systems without permission.
• Remove notices, hide attribution, or misrepresent FildraAI outputs as independently verified professional advice.

6.3 Generated Outputs

You may use generated outputs for your authorized personal, farm, research, internal business, or operational purposes, but you acknowledge that:

• Outputs may be incomplete, uncertain, outdated, or incorrect.
• Similar or identical outputs may be generated for other users with similar inputs.
• You should not present outputs as guaranteed, certified, professional, legal, medical, financial, or regulatory advice.
• You are responsible for reviewing outputs before relying on or sharing them.

7. Emergency and High-Risk Situations

Seek professional or official help immediately for situations including:

• Medical, veterinary, poisoning, contamination, or life-threatening situations.
• Regulatory, quarantine, legal, certification, or movement-restriction decisions.
• Major financial, contractual, employment, or safety decisions.
• Security incidents, suspected account compromise, leaked API keys, or unauthorized access.
• Critical operational failures where delay may cause serious harm or loss.

8. Enforcement and Consequences

8.1 Violation Handling

We may investigate and act on violations of this policy, our Terms, security requirements, or applicable law. Depending on severity, we may:

• Issue warnings or request corrective action for minor issues.
• Rate limit, restrict, revoke, or rotate API keys and access tokens.
• Temporarily suspend features, accounts, workspaces, or integrations during investigation.
• Terminate accounts for serious, repeated, abusive, illegal, or harmful violations.
• Preserve evidence, seek legal remedies, or report illegal activity to appropriate authorities.

8.2 Appeals

If you believe an enforcement action was made in error, you may contact us with:

• The account, workspace, API key, or email address involved.
• A clear explanation of the activity and why you believe the action was incorrect.
• Any steps you have taken to prevent future issues, such as securing credentials or adjusting traffic patterns.

9. Help Us Improve Responsible Use

Responsible use is something we build together. Your feedback helps us improve safety, reliability, accessibility, and clarity across FildraAI services:

• Report outputs that appear unsafe, inaccurate, misleading, or inconsistent with local reality.
• Tell us when translations, transcriptions, or summaries fail important meaning.
• Report confusing UI flows, missing warnings, or unclear consent/permission behavior.
• Notify us if you believe your account or API key has been compromised.
• Suggest improvements to documentation, warnings, service limits, or accessibility.