Acceptable Use Policy

Last updated: 3 August 2026

This Acceptable Use Policy defines acceptable behavior on MaidServSA and explains how violations are reported and handled.

Table of Contents

• 1. Purpose and Scope
• 2. Acceptable Use
• 3. Prohibited Activities
• 4. Content Standards
• 5. Reporting Violations
• 6. Enforcement
• 7. AI Use Policy

1. Purpose and Scope

This policy applies to all users and all platform surfaces, including profiles, job posts, applications, messages, ratings, and uploaded files.

2. Acceptable Use

You must:

• Use MaidServSA only for lawful domestic work marketplace activity.
• Provide accurate profile and job information.
• Communicate respectfully and professionally.
• Submit fair and honest ratings/reviews.
• Respect privacy and consent boundaries.

3. Prohibited Activities

You may not:

• Commit fraud, impersonation, or identity misuse.
• Harass, threaten, stalk, or intimidate others.
• Post discriminatory, hateful, or abusive content.
• Submit false, malicious, or retaliatory ratings/reports.
• Attempt scraping, automated extraction, or unauthorized access.
• Upload malware or harmful files.
• Share another person's personal information without authorization.
• Circumvent platform safeguards or moderation decisions.

4. Content Standards

Content must be:

• Accurate and not misleading.
• Relevant to the purpose of the platform.
• Respectful and lawful.
• Free from sexual exploitation, threats, extortion, or hate speech.

For ratings and reviews:

• Keep feedback factual and job-related.
• Do not use ratings to blackmail or pressure other users.

4.1 Child safety and CSAE

MaidServSA explicitly prohibits child sexual abuse and exploitation (CSAE). Users may not use MaidServSA to groom, sexually exploit, sexually solicit, sextort, traffic, or otherwise endanger a child. Users may not create, request, upload, store, link to, distribute, or knowingly facilitate child sexual abuse material (CSAM), including visual material involving a minor in sexually explicit conduct.

These restrictions apply to profiles, job posts, applications, messages, ratings, uploaded files, and any other MaidServSA surface. They apply even though MaidServSA is intended for users aged 18 and older.

If MaidServSA obtains actual knowledge of CSAM or CSAE activity, we will take appropriate action under this policy and applicable law. This may include restricting or removing the content, preserving relevant information as legally permitted, suspending or banning accounts, and referring the matter to law enforcement or the applicable reporting authority where required.

5. Reporting Violations

You can report policy violations directly in the app using the Report button on the user profile, job, application, message, or rating you are concerned about. When you report something, you choose a reason and a severity level (low, medium, or high), and you can add extra detail if you wish.

You can report:

• User profiles, jobs, applications, messages, or suspicious activity.
• A specific message in a conversation.
• A specific rating or review.

For an urgent child-safety concern, email support@maidservsa.co.za directly with CSAE/CSAM in the subject line — do not wait to find or use an in-app report option first. If the in-app report flow is available on your device, selecting the highest severity available also helps our team triage, but it is never a required step before contacting support. Do not attach or forward suspected CSAM unless an authorized investigator specifically requests it.

Our admin team reviews every report and takes appropriate action. For urgent concerns, you can also email support@maidservsa.co.za.

6. Enforcement

When a report is submitted, the reported item (job, application, or message) is automatically locked and cannot be edited or deleted while the report is open or under review. The lock is released when an admin resolves the report.

Admins may apply one of the following actions when resolving a general report:

For message and rating reports, additional actions include message redaction and rating hiding/restoration.

Enforcement decisions consider severity, repetition, and harm risk. Escalation to legal authorities may follow serious misconduct.

7. AI Use Policy

MaidServSA offers optional AI assistance to help clients and maids work more efficiently. This section is the authoritative compliance reference for all AI features on the platform. AI assistance is never required to use MaidServSA — every page works without it.

7.1 AI features in scope

7.2 Preview usage quota

While the platform is in preview, every user — CLIENT, MAID, and ADMIN — receives 5 AI calls in total across all four AI features. The quota is enforced on our servers and cannot be bypassed.

• A welcome modal explains the quota on first signup and on the first login of each day.
• A live counter ("n/5 AI uses left") is shown inside every AI button.
• When the quota reaches 0 the buttons remain visible but disabled. A one-time per-session modal explains that subscription tiers will follow.
• A failed AI call (model error, validation rejection, no recommendations possible) is refunded automatically — your quota is only spent on successful generations.
• 0-applicant requests for the best-candidate picker return a friendly error with no quota spend. 1-applicant requests short-circuit to that maid with no model call and no quota spend.

Subscription tiers will be introduced before the platform leaves preview. This policy will be updated when paid usage begins.

7.3 Third-party AI processor

AI generations are produced by Google's Gemini AI service. Google is our sub-processor for these features. Prompt content sent to Google is processed under the Google Cloud Generative AI terms and may be processed in regions outside South Africa.

Cross-border processing safeguards are described in § 11 of the Privacy Policy and § 6.2.

7.4 What data is sent to the AI model

The model receives only the fields strictly needed to produce the requested output:

7.5 What is never sent to the AI model

The following data is never transmitted to Google Gemini under any AI feature:

• Email addresses, phone numbers, physical addresses, or place IDs.
• Exact GPS coordinates.
• Date of birth, ID document numbers, ID images, or selfies.
• Passwords, password hashes, password-reset tokens, or auth tokens.
• Payment or banking details (the platform does not handle these in any case).
• Direct-message content between users.
• CV file contents or any other uploaded file bytes — only the filename string, if any, is used.
• Identity verification records.

7.6 Prompt injection and untrusted content

User-supplied free-text (bios, application messages, experience answers, special requirements) is treated as untrusted job data, not instructions. The prompt scaffolding explicitly frames such text as content to summarise, not directives to obey. Outputs are additionally validated:

• The best-candidate picker constrains the AI to a fixed response format.
• Our servers check that the recommended candidate belongs to the actual applicant set; any mismatch is rejected and the quota refunded.

7.7 Accuracy, bias, and informational nature

AI outputs are suggestions only. They:

• Are not professional, legal, employment, or safety advice.
• May contain factual errors, hallucinated details, or omissions.
• May reflect biases present in the underlying model.
• Do not replace the user's own judgement when posting jobs, applying, or selecting a maid.

The best-candidate picker is a ranking aid; the final hiring decision and all responsibility for it rest with the CLIENT. Selecting an AI-recommended candidate does not constitute a warranty of fitness, conduct, or skill.

7.8 User responsibility for AI-assisted content

Anything you publish on the platform is your content, even when an AI feature drafted it. You must review every AI draft before saving or sending and ensure it is:

• Truthful and not misleading.
• Non-discriminatory and non-harassing.
• Non-defamatory and lawful.
• Compliant with this policy, the Terms of Service, and applicable South African law.

You may not use AI features to:

• Generate fraudulent listings, fake applications, or impersonate another person.
• Produce content that targets a protected class, incites violence, or sexualises any person.
• Probe the model with adversarial prompts (jailbreaks, prompt injection, data-exfiltration attempts).
• Automate scraping of AI outputs or resell AI generations.

Violations are subject to enforcement under § 6, including suspension and ban.

7.9 Opt-out and the right not to use AI

AI features are strictly optional:

• All forms accept manual input — every AI button has a non-AI equivalent path.
• Choosing not to click AI buttons is the opt-out; no AI processing occurs unless you trigger it.
• AI features are gated behind your authenticated request, never invoked on background data.

7.10 Retention and logging of AI activity

• Our AI-usage records store only your user ID, the AI uses you have remaining, the total you have used, and the relevant timestamps. They do not store prompts or AI outputs.
• Server-side request logs may capture timing, outcome, and quota state for security and abuse-prevention purposes; they do not persist full prompts or outputs.
• Final accepted AI-drafted content (job descriptions, application messages) is stored as ordinary user content in the database under the regular Data Retention Policy.
• Google's retention of prompts and outputs follows the Google Cloud Generative AI terms; see Google's documentation for current commitments.

7.11 Refunds, errors, and reporting

• If an AI call fails after consuming a token, the server refunds the token automatically.
• If you believe an AI output was harmful, offensive, deceptive, or unsafe, report it via the in-app report flow described in § 5 or email support@maidservsa.co.za with subject line AI Output Concern.
• Concerns about how your personal data is handled in AI features can be raised under the Privacy Policy § 7 Data Subject Rights.

7.12 Children, sensitive personal information, and special categories

• The platform is for users 18+ (Terms § 4.1). Do not include any minor's personal information in AI prompts.
• Avoid putting special personal information (health, religion, political views, biometrics) into AI-bound fields — none of the AI features require it, and content you submit may be processed by a third party.

7.13 Changes to this policy

This policy may evolve as new AI features ship or as subscription tiers replace the preview quota. The "Last updated" date at the top reflects the most recent material change. Substantive changes will also appear in the Changelog.

Related policies:

• Terms of Service — § 8 Intellectual Property and AI Content
• Privacy Policy — § 6.2 Service providers and § 11 Cross-Border Transfers
• Intellectual Property Policy
• Data Retention Policy
• Dispute Resolution
• Safety Guidelines