AI Usage Policy

1. Purpose & Scope

At Contingent Works, we use Artificial Intelligence (AI) to enhance the efficiency and quality of our operations. AI is a tool to assist and optimise our work processes, but it does not replace human judgement. All externally shared outputs receive human review to ensure quality and accuracy. This policy outlines the guidelines for the use of AI across our organisation, applicable to all employees and contractors.

The scope of this policy includes the use of AI in content creation, data processing, and decision-making processes within Contingent Works. It covers all AI-related activities, ensuring compliance with ethical standards and data protection regulations.

2. Definitions

AI (Artificial Intelligence): A branch of computer science focused on creating systems capable of performing tasks that typically require human intelligence, such as decision-making, language translation, and visual perception.

  • Generative AI: A type of AI that can generate new content, such as text, images, or code, based on training data.

  • LLM (Large Language Model): A type of AI model designed to understand and generate human language, such as ChatGPT.

  • Closed vs Open Tools: Closed tools are proprietary systems with restricted access, while Open tools are publicly accessible and modifiable.

3. Guiding Principles

Contingent Works follows key guiding principles in using AI:

  1. Human-in-the-Loop: Significant decisions and all public/client-facing outputs require human review prior to use or publication.

  2. Transparency: We clearly disclose the use of AI assistance where material, ensuring stakeholders understand when and how AI contributes to outputs.

  3. Fairness: AI tools and processes are regularly evaluated for bias and fairness, with staff trained to detect and mitigate these issues.

  4. Safety: We prioritise the safety and integrity of our data and outputs, ensuring compliance with organisational standards.

4. Data Protection & Classification Rules

We classify data into four levels with specific rules:

  1. Restricted: Includes special category data and trade secrets. These must never be input into public AI tools. Use internal models with strict access controls and anonymise data where feasible.

  2. Confidential: Covers client drafts and internal strategies. Use approved enterprise tools with data controls; human review is mandatory before external sharing.

  3. Internal: Includes process documents and internal announcements. Approved tools may be used, but outputs must be spot-checked for hallucinations and bias.

  4. Public: Encompasses published marketing and research. Approved tools can be used, but factual claims and sources must be verified.

In line with these classifications, do not input confidential, personal, or client-identifying information into public or unapproved AI tools.

5. Permitted Uses

Within Contingent Works, AI can be used for:

  1. Brainstorming and Outline Generation: AI can assist in generating ideas and structuring documents.

  2. Editing, Summarisation, Translation: These tasks can be performed by AI, but human review is required to ensure accuracy and appropriateness.

  3. Drafting Non-Sensitive Communications: Routine emails and internal messages can be drafted with AI assistance.

  4. Code Refactoring: AI may be used to refactor code on non-proprietary snippets, enhancing efficiency and readability.

6. Prohibited Uses

To protect our business and clients, certain uses of AI are prohibited:

  1. Uploading Client/Customer PII: Personal Identifiable Information should never be input into public AI tools.

  2. Relying on AI for Formal Advice: AI-generated outputs should not replace expert review in legal, medical, or financial contexts.

  3. Deepfakes and Impersonations: Creating misleading representations using AI is strictly forbidden.

  4. Using Unapproved AI Tools: Only tools on the approved list may be used for company work.

7. Approved Tools & Access Control

Contingent Works only permits the use of the following AI tools:

  1. ChatGPT

  2. Google Gemini

  3. Microsoft Copilot

Employees must use these tools for business purposes only. Any requests for new tools must follow the Tool Approval Workflow:

1. Submit Request: Employee submits a tool request form. 2. Manager Review: People Manager reviews the request for relevance and necessity. 3. Security Assessment: AI Oversight Lead conducts a security and compliance assessment.

8. Governance & Roles

  1. AI Oversight Lead / Manager: Responsible for maintaining this policy, managing the approved tools list, handling exceptions, and coordinating training and audits.

  2. People Managers: Approve team use cases and ensure compliance with AI usage guidelines, providing coaching as needed.

  3. All Staff: Must adhere to data classification rules, review AI outputs, and report issues promptly.

9. Sector-Specific Addenda

In the general business sector, Contingent Works recognises the importance of AI in enhancing productivity and content quality. For example, AI can be employed to draft marketing materials, with a focus on maintaining creativity and factual integrity through human oversight.

10. Training & Literacy

Training is essential to ensure responsible AI use. Our training programme covers:

  1. AI Basics and Limitations: Understanding what AI can and cannot do.

  2. Data Classification and Safe Prompting: Proper handling of data when using AI tools.

  3. Bias Detection and Inclusive Review: Techniques for identifying and addressing bias.

  4. Disclosure of AI Assistance: When and how to inform stakeholders of AI involvement.

  5. Approved Tool List & Request Process: Familiarity with approved tools and how to request new ones.

11. Incident & Exception Handling

Prompt handling of incidents is crucial:

  1. Triggers: Include the input of PII into unapproved tools, use of harmful or biased outputs externally, or copyright complaints on AI-assisted work.

  2. Actions: Immediate containment, notify the AI Oversight Lead, inform the Data Protection Officer if personal data is involved, conduct a root cause analysis, update policies, and provide refresher training if systemic issues are identified.

12. Review & Updates

This AI Usage Policy will be reviewed annually or when significant changes in technology, regulation, or business practices occur. Version updates will be documented, with the latest version always available to all staff.

Version Control:

  1. Version 1.0: Initial release, December 2025

This comprehensive policy ensures that Contingent Works utilises AI effectively while maintaining ethical standards and protecting data integrity.