Effective Date: 2025-04-17

1. Introduction

Nutrish.ai actively explores how AI can enhance our offering for consumers and B2B partners by delivering personalised nutrition support and improving operational efficiency. We utilise AI and automation to generate dietary recommendations, create tailored meal plans, and provide conversational support through chat-based interfaces. As powerful generative AI models become more accessible, we recognise both the potential and the risks associated with them—including legal, ethical, and user safety concerns.

These are company-wide guidelines for the use of generative AI and AI-generated content at Nutrish.ai. The policy applies across the business and will be continuously updated as new use cases, risks, and regulations emerge. This first version focuses on user-facing content and internal use of generative AI. 

2. Definitions

Generative AI refers to technologies that can create new, unique content (e.g. text, images, audio, code) based on training data.
Artificial Intelligence is a broad term for machine-based systems capable of producing outputs (e.g. predictions, recommendations, decisions) intended to achieve specific objectives.
AI-generated content is any output produced by an AI system with minimal or no human intervention.

3. Guidelines for Generative AI at Nutrish.ai

All AI-generated content made available to users must follow the same ethical, factual, and safety standards as human-generated content. Nutrition-related content must be rooted in scientific evidence and aligned with best practices.

• Any direct use of generative AI in content exposed to end-users must be approved by the responsible domain owner (e.g. Nutritionist or Product Owner). Human review is mandatory before publication to ensure accuracy, relevance, and safety.
  
  

• All public-facing AI-generated content must be transparently labelled in line with current recommendations. For example, a chatbot must introduce itself as an AI assistant, and meal plans generated by AI must indicate this clearly in context.
  
  

• Any AI-generated content that may influence user behaviour or health decisions must be carefully reviewed through regular quality assurance processes and, where appropriate, co-developed or validated by qualified professionals.
  
  

• Persistent or large-scale use of generative AI (e.g. as part of product features, marketing automation, or user interfaces) must be approved by the Executive Team or a designated decision-maker. Use of synthetic content such as AI-generated voices or images in product interfaces requires heightened scrutiny.
  
  

• Nutrish.ai must remain vigilant regarding external AI-generated content, including misinformation, deepfakes, or manipulated material. When referencing external sources, factual validation is required.
  
  

• Biases may be embedded in AI models and data. Content must be assessed for neutrality, inclusiveness, and cultural relevance. Measures should be taken to avoid reinforcing harmful stereotypes or disinformation.
  
  

• Each department using generative AI must ensure a clear understanding of rights and obligations for AI-generated content, including usage terms and intellectual property rights outlined in service agreements.
  
  

• Sensitive company information or user data must never be shared in AI prompts using third-party services. Anything entered in a public AI tool is to be treated as public. If third-party data is included in prompts, the team is responsible for verifying that rights and legal bases exist.
  
  

• Nutrish.ai emails may not be used to create accounts in third-party AI services unless security and privacy have been explicitly assessed and approved.
  
  

4. Checklist for Public-Facing AI Use

• Could the use of AI-generated content affect user trust or decision-making?

• Has the responsible lead reviewed and approved the content?

• What is known about the model’s training data? Are there risks of bias?

• Can we guarantee that no rights or obligations are being violated?

• Have we applied appropriate labelling and disclosure for AI-generated content?

5. Checklist for Internal AI Use

• Does the prompt contain sensitive company or personal information?

• Do we have rights to share the input or reuse the output?

• Have outputs been validated for factual accuracy and brand safety?

• Have you discussed the use case with your manager or team lead?
  
  

6. Contact and Responsibility

For questions or guidance, contact the Nutrish.ai AI Working Group:
email@nutrish.ai
Policy ownership: Executive Leadership Team
Policy oversight: AI Working Group