AI Policy

COMPANY: APEX24 TECHSYS AND BRAND: BUCKLEUP - AI GOVERNANCE & RESPONSIBLE AI POLICY

Effective Date: 1 FEBRUARY 2026

1. PURPOSE

This AI Governance & Responsible AI Policy (“AI Policy”) establishes the principles, controls, and governance framework under which BuckleUp designs, deploys, and operates artificial intelligence systems, including the AI Photo Product and any AI-enabled educational tools.

This Policy applies to:

• All AI models deployed within the BuckleUp ecosystem
• All AI-based image processing systems
• All automated decision-support systems
• All third-party AI integration

The purpose of this Policy is to:

• Protect minors and vulnerable individuals
• Ensure lawful and ethical AI use
• Align with Indian law including the DPDP Act 2023
• Mitigate operational, reputational, and regulatory risk
• Provide transparency to institutions and parents

2. SCOPE

This Policy applies to:

• Schools using BuckleUp
• Parents and guardians
• Teachers and administrators
• Internal employees
• Contractors
• AI vendors and subprocessors

It governs:

• Model selection
• Model deployment
• Data processing
• Monitoring
• Risk mitigation
• Incident response

3. CORE AI PRINCIPLES

BuckleUp’s AI systems operate under the following principles:

3.1 Child Safety First

AI systems must not:

• Exploit minors
• Profile children for behavioral advertising
• Infer psychological traits
• Enable surveillance

3.2 Human-Centered Design

AI outputs are assistive, not authoritative.
Human oversight remains mandatory in institutional contexts.

3.3 Transparency

Users must understand:

• When AI is being used
• What type of processing occurs
• What limitations exist

3.4 Proportionality

AI systems must process only the minimum data necessary to achieve stated purposes.

3.5 Security by Design

Security controls must be integrated at:

• Model deployment level
• Data storage level
• API integration level

4. AI SYSTEM CATEGORIES

BuckleUp classifies AI systems into the following internal risk categories:

4.1 Low Risk AI

• Lighting correction
• Color balancing
• Background blur
• Artistic filters

4.2 Moderate Risk AI

• Face detection for enhancement
• Group image alignment
• Style transformation

4.3 Prohibited AI Use Cases

The following are strictly prohibited:

• Facial recognition for identity tracking
• Emotion inference for behavioral assessment
• Social scoring of students
• Political persuasion targeting minors
• Deepfake impersonation tools

5. DATA GOVERNANCE FRAMEWORK

5.1 Data Minimization

AI models shall only process:

• Required image data
• Limited metadata
• Non-sensitive contextual information

5.2 Biometric Safeguards

• No persistent biometric database maintained
• No cross-platform facial matching
• No third-party biometric sale or sharing

5.3 Model Training Restrictions

Uploaded student images:

• Are not used to train public AI models
• Are not incorporated into foundation models
• Are not reused outside stated service scope without explicit consent

6. MODEL DEVELOPMENT & PROCUREMENT

6.1 Internal Model Controls

Where BuckleUp develops internal models:

• Model architecture documentation is maintained
• Training data sources are documented
• Risk assessment is conducted pre-deployment
• Bias evaluation is performed

6.2 Third-Party AI Vendors

Where third-party AI services are used:

• Vendor due diligence is conducted
• Data processing agreements are executed
• Vendor security posture is evaluated
• Vendor compliance with Indian law is assessed

7. AI RISK MANAGEMENT FRAMEWORK

BuckleUp maintains a structured AI Risk Framework:

7.1 Pre-Deployment Risk Assessment

Before deploying new AI features:

• Data sensitivity classification
• Minor data exposure analysis
• Ethical impact review
• Legal compliance validation

7.2 Ongoing Monitoring

• Model performance monitoring
• Safety filter review
• Incident reporting mechanism
• Periodic governance review

8. FAIRNESS & BIAS MITIGATION

AI systems must not:

• Discriminate based on race, religion, caste, gender, disability, or socio-economic status
• Apply differential processing across demographic groups

Bias detection processes include:

• Sample testing
• Model output audits
• User feedback analysis

9. HUMAN OVERSIGHT

AI outputs are:

• Non-binding
• Non-decisional
• Subject to human review

Teachers and institutions must evaluate AI-generated images before distribution.

10. SECURITY CONTROLS FOR AI SYSTEMS

Security measures include:

• Encrypted API communications
• Access token restrictions
• Role-based access
• Monitoring for abnormal usage
• Abuse detection systems

11. INCIDENT RESPONSE

In case of AI-related incidents (including harmful output or misuse):

1. Incident is logged
2. Impact assessment conducted
3. Feature may be temporarily suspended
4. Affected institutions notified where required
5. Root cause analysis conducted

12. ACCOUNTABILITY STRUCTURE

12.1 AI Governance Lead

A designated AI Governance Lead oversees:

• Compliance
• Risk evaluation
• Vendor assessments

12.2 Executive Oversight

AI governance is reviewed periodically by:

• Senior leadership
• Legal advisors
• Security teams

13. USER RESPONSIBILITIES

Institutions and parents must:

• Avoid misuse of AI features
• Avoid uploading non-consensual images
• Report suspicious activity

14. LIMITATIONS OF AI SYSTEMS

AI systems:

• Operate probabilistically
• May generate imperfect outputs
• May require manual correction

No guarantee is made regarding artistic satisfaction.

15. PROHIBITED USE POLICY

Users may not use BuckleUp AI systems for:

• Harassment
• Misinformation
• Political propaganda
• Identity impersonation
• Fraud
• Academic manipulation

16. REGULATORY ALIGNMENT

This AI Policy aligns with:

• Digital Personal Data Protection Act, 2023
• Indian IT Act, 2000
• Child safety principles
• Enterprise SaaS governance standards

17. AUDIT & REVIEW

This AI Policy shall be:

• Reviewed annually
• Updated as AI capabilities evolve
• Updated in response to regulatory changes

18. TRANSPARENCY TO USERS

Users are informed when:

• AI processing is applied
• New AI features are introduced
• Significant changes occur

19. CROSS-BORDER AI PROCESSING

Where AI infrastructure operates outside India:

• Safeguards are applied
• Contractual protections are implemented
• Data localization requirements are respected where applicable

20. FUTURE AI CAPABILITIES

If BuckleUp expands AI capabilities beyond photo processing (e.g., academic analytics), a separate AI Impact Assessment shall be conducted prior to deployment.

21. ENFORCEMENT

Violation of this Policy may result in:

• Account suspension
• Institutional notification
• Contract termination

22. CONTACT FOR AI GOVERNANCE

AI Governance Queries:
Email: contact@buckleupnow.in