Tonita Taylor Consulting AI Usage Policy

1. Purpose of This Policy

This policy sets out how Tonita Taylor Consulting uses Artificial Intelligence (AI) in a responsible, ethical and transparent manner. It outlines our principles, safeguards and procedures to ensure AI enhances our services without compromising privacy, safety or trust.

We align with the Australian Government’s Voluntary AI Safety Standard Guardrails (2024), the Australian Privacy Principles (APPs), and relevant professional codes in research and evaluation.

2. Scope and Definition of AI

This policy applies to all AI tools and systems used by Tonita Taylor Consulting, whether internally or in client-facing projects. AI tools and systems may include:

• Machine learning algorithms
• Natural Language Processing (NLP) tools
• Generative AI (e.g. large language models)
• AI-powered transcription, translation or analysis tools

AI is defined here as any system that performs tasks typically requiring human intelligence, including learning from data, generating content, or making predictions.

3. Our Approach to Responsible AI

We use AI to support, not replace, human expertise. As a general rule, AI may be used to perform the functions of a junior researcher, analyst or evaluator. All AI-generated work is subject to human review for accuracy, completeness, and bias.  In most cases, all work is edited, re-worked and/or adapted by a human.

We do not use AI to make decisions. All decisions, whether strategic, evaluative, or operational, are made by qualified humans. AI may assist with information processing, but it will never be the final arbiter.

Our approach is guided by the following principles:

• Human-centred design
• Ethical alignment
• Transparency
• Privacy and data protection
• Accountability
• Continuous improvement

4. Accountability for AI Use

The Chief Evaluator is responsible for AI governance. Each AI tool or process is overseen by a human who ensures it is used appropriately and that outputs are reviewed before use.

We maintain an internal register of AI tools, including their purpose, configuration, and responsible personnel. All decisions involving AI are documented and subject to review.

5. Risk Management and Impact Assessment

Before deploying any AI tool, we assess risks including:

• Accuracy and reliability
• Bias or discrimination
• Privacy and data protection
• Impact on stakeholders
• Ethical and reputational risks

We mitigate risks through pilot testing, human oversight, data minimisation, and fallback procedures. Risk assessments are documented and reviewed periodically.

6. Data Protection and Privacy

We treat data protection as paramount in all AI use. Specifically:

• We will only utilise AI tools and platforms where a suitable enterprise agreement is in place, guaranteeing robust data protection and compliance with our privacy obligations. Typically, Tonita Taylor Consulting uses the licenced enterprise version of Microsoft Copilot. This means that no client data is ever shared with public AI models or used to train external systems. All data remains securely within our Microsoft 365 environment and is protected by enterprise-grade privacy, security and compliance standards.
• Other, more appropriate AI tools and platforms may be used from time to time, however we will never use any AI tool that retains, stores or trains on client data unless we have contractual guarantees that the data will not be used for model training or any purpose beyond our control.
• We avoid AI platforms that use input data to improve their models unless we have an enterprise agreement that explicitly prohibits such use.
• All AI use complies with our Privacy Policy and the APPs.
• We anonymise or de-identify data wherever possible before AI processing.
• We do not use AI to scrape or collect personal data from external sources without consent.

If an AI tool requires cloud processing, we ensure the provider meets Australian privacy standards or equivalent protections. We do not use tools that store data in jurisdictions without adequate safeguards.

7. Transparency and Disclosure

We are open about our use of AI. We will:

• Inform clients when AI is used in their projects
• Label AI-generated content where appropriate
• Explain how AI outputs were derived
• Provide human review and interpretation of AI results
• Offer alternatives to AI interaction where feasible

We publish this policy and update it as needed. Stakeholders may request further details about our AI tools and processes.

8. Human Oversight and Control

AI tools are always supervised by humans. Oversight includes:

• Reviewing AI outputs before use
• Maintaining manual alternatives
• Retaining the ability to override or disable AI systems
• Ensuring that individuals affected by AI outputs can request human review

All AI-generated work is treated as draft or preliminary until verified by a qualified human.

9. Testing and Reliability

We test AI tools before deployment and monitor them continuously. Testing includes:

• Accuracy checks against known data
• Bias audits
• Edge case testing
• Performance monitoring
• Documentation of results and adjustments

We do not rely on AI tools that fail to meet our standards for reliability and fairness.

10. Monitoring and Evaluation

We monitor AI use during projects and evaluate its impact regularly. This includes:

• Logging errors or anomalies
• Gathering user and stakeholder feedback
• Reviewing performance metrics
• Updating risk assessments
• Adjusting usage or discontinuing tools if needed.

11. Record-Keeping and Documentation

We maintain records of:

• AI tools and configurations
• Risk assessments and approvals
• Inputs and outputs
• Decisions influenced by AI
• Incidents and resolutions

These records support accountability and auditability.

12. Training and Awareness

All staff receive training on:

• AI ethics and safety
• Tool usage and limitations
• Privacy and data handling
• Incident response procedures

We encourage continuous learning and share knowledge internally.

13. Incident Response and Review

If an AI-related incident occurs, we will:

• Contain and investigate the issue
• Notify affected parties and regulators if required
• Resolve the problem and document actions
• Review root causes and update procedures
• Communicate outcomes transparently.

14. Review and Updates to This Policy

We will review and update this policy to reflect changes in law or practice. While significant changes may be communicated directly with current and/or past clients, all changes will be reflected in the updated policy on the website.  We commit to remaining aligned with evolving standards and regulations. This Policy is current as at October 2025.

15. Contact Details

For questions, feedback or concerns about this policy:

Chief Evaluator: Tonita Taylor
Email: Web contact form
Phone: +61 0413 565 052