Seller: Superspeed.ai Pty Ltd
IP Owner & Licensor: Strategic Global Holdings Pty Ltd (ACN 693 256 503)
Effective Date: 1 January 2025
Version: 4.95 Ultra-Final
Document Owner: Chief AI Governance Officer (CAGO), Superspeed.ai Pty Ltd
Review Cycle: Annual or upon release of updated AI regulations or standards
1. Definitions
Includes: AI System, Machine Learning, Automated Decision-Making (ADM), Profiling, Model Drift, Training Data, Inference, AI-Assisted Output, High-Risk Processing, Human-in-the-Loop, and Explainability.
2. Purpose & Scope
This Statement outlines how AI systems are governed, managed, monitored, and reviewed across the bookstore, platform recommendations, customer support tools, search optimisation, and fraud-monitoring systems.
3. AI Governance Principles
Superspeed.ai adopts governance aligned with:
- OECD AI Principles
- ISO/IEC 42001 (AI Management Systems)
- EU AI Act (non-binding but anticipatory compliance)
- NIST AI Risk Management Framework
Key principles include fairness, transparency, accountability, safety, privacy, human oversight, and data minimisation.
4. AI Use Cases on the Platform
AI is used for:
- Personalised book recommendations
- Search ranking optimisation
- Fraud detection and anomaly monitoring
- Automated support suggestions
AI is not used for:
- Credit decisions
- Legal determinations
- Employment or eligibility screening
- Any high-risk or consequential automated decisions
5. Human Oversight (Human-in-the-Loop)
All AI-assisted outputs are subject to human review. Staff may override any AI suggestion. Critical decisions (refunds, account restrictions, escalations) require direct human intervention.
6. Data Governance & Protection
AI systems operate in accordance with the Global Privacy Policy, GDPR/UK Addendum, CCPA/CPRA Addendum, PIPL Addendum, PDPA Addendum, and the Cookie Policies. No Sensitive Personal Data is used in AI models.
7. Training Data Controls
Training data is curated to:
- Exclude discriminatory or biased sources
- Maintain data minimisation
- Prevent ingestion of copyrighted Digital Content unless explicitly permitted
Digital books or content sold on the platform are never used for training.
8. Explainability & Transparency
Users may request explanation of AI-assisted interactions. Documentation is maintained describing:
- How models influence recommendations
- When AI is used vs not used
- Which data categories AI relies on
9. Bias Detection & Mitigation
We perform periodic assessments to detect and reduce bias using:
- Model drift detection
- Fairness assessments
- Diverse evaluation datasets
- Manual review of flagged cases
10. Monitoring & Quality Assurance
AI systems undergo:
- Continuous monitoring
- Error rate analysis
- Performance audits
- Regular versioning reviews
- Controlled deployment with rollback capability
11. Incident Response (AI-Specific)
AI incidents—such as unexpected model behaviour, fairness failures, or degraded accuracy—trigger:
- Isolation or rollback of the model
- Human review and override
- Root-cause investigation
- User notification where applicable
12. Prohibited AI Uses
Superspeed.ai prohibits:
- AI systems that make legally binding decisions without human oversight
- Use of biometric or facial recognition data
- Emotion inference, predictive profiling, or high-risk biometric processing
- Training AI on customer-owned or licensed Digital Content
13. Cross-Document Integration
This Statement aligns with: AI Explainability & Human Oversight Notice, Automated Decision-Making & Profiling Disclosure, Privacy Policy, Security Overview, Website Terms of Use, and Business Continuity Statement.