AI Transparency & Automated Decision-Making

Effective Date: January 3, 2025

Last Updated: August 3, 2025

AI Development Status

This transparency statement describes our planned AI systems and decision-making processes. AI features are currently in development. Some technical capabilities described may not yet be fully implemented. Contact info@reclaimed.health for current AI feature availability.

Our Commitment to AI Transparency

Reclaimed Health Ltd uses artificial intelligence (AI) to help UK families make safer food choices for their children. We believe in being completely transparent about how our AI systems work, their limitations, and how they make decisions that could affect your family’s health and safety.

This AI Transparency Statement explains our automated decision-making processes in accordance with UK GDPR Article 22, the EU AI Act, and UK ethical AI principles.

How We Use AI in Food Safety

AI-Powered Features

🔍 Ingredient Analysis

What it does: Analyses food ingredient lists to identify potential health concerns

How it works: Natural language processing to parse ingredient lists and cross-reference with safety databases

Human oversight: All high-risk ingredients flagged for human expert review

⚠️ Allergy Detection

What it does: Identifies allergens and cross-contamination risks in food products

How it works: Pattern matching against comprehensive allergen databases and regulatory lists

Human oversight: Critical allergy warnings always verified by food safety experts

🏥 Health Risk Assessment

What it does: Evaluates potential health impacts of food ingredients

How it works: Machine learning models trained on peer-reviewed research and regulatory data

Human oversight: All health recommendations reviewed by nutrition professionals

📱 Personalised Recommendations

What it does: Suggests safer food alternatives based on your family’s preferences

How it works: Recommendation algorithms considering dietary restrictions, allergies, and preferences

Human oversight: Recommendations filtered by nutritionists for appropriateness

AI Decision-Making Process

How Our AI Makes Decisions

  1. Data Input: You scan a food product or enter ingredient information
  2. AI Analysis: Our algorithms analyse ingredients against safety databases
  3. Risk Calculation: AI calculates potential health risks based on current research
  4. Human Review: High-risk results reviewed by food safety experts
  5. Recommendation: Personalised advice provided with confidence levels
  6. Continuous Learning: System learns from new research and user feedback

Decision Factors

Our AI considers multiple factors when making recommendations:

  • Regulatory Status: UK FSA, EFSA, and international safety assessments
  • Scientific Research: Peer-reviewed studies on ingredient safety
  • Population Data: Age-specific safety guidelines (especially for children)
  • Individual Factors: Personal allergies, dietary restrictions, health conditions
  • Dosage Levels: Concentration and typical consumption amounts
  • Cumulative Exposure: Combined exposure from multiple food sources

AI Training Data and Sources

Training Data Sources

Regulatory Databases

UK FSA, EFSA, FDA, Health Canada food safety databases and regulatory decisions

Scientific Literature

Peer-reviewed research from PubMed, nutrition journals, and toxicology studies

Commercial Food Databases

Publicly available ingredient lists and nutritional information from major food manufacturers

Expert Knowledge

Input from registered nutritionists, food scientists, and paediatric health experts

Data Quality and Bias Prevention

  • Diverse Sources: Training data from multiple geographical regions and populations
  • Regular Updates: Models retrained monthly with latest research and regulatory changes
  • Bias Testing: Regular testing for demographic, cultural, and socioeconomic biases
  • Expert Review: All training data validated by qualified food safety professionals
  • Transparency: Clear documentation of data sources and model limitations

AI Accuracy and Limitations

Current Accuracy Levels

Allergen Detection: 98.5% accuracy (verified against known allergen databases)

Ingredient Classification: 94.2% accuracy (for common food ingredients)

Health Risk Assessment: 89.7% agreement with expert nutritionist reviews

Recommendation Relevance: 91.3% user satisfaction rating

Note: Accuracy rates updated monthly based on continuous testing and validation

Known Limitations

Important Limitations:

  • Novel Ingredients: Lower accuracy for new or rare food ingredients not well-represented in training data
  • Complex Interactions: Cannot predict all possible ingredient interactions or cumulative effects
  • Individual Variations: Cannot account for unique individual sensitivities or medical conditions
  • Processing Effects: Limited understanding of how food processing affects ingredient safety
  • Regional Variations: Some training data may not reflect UK-specific dietary patterns
  • Emerging Research: May not immediately incorporate very recent scientific findings

Confidence Levels

All AI recommendations include confidence levels:

  • High Confidence (90-100%): Well-established safety data, regulatory consensus
  • Medium Confidence (70-89%): Good evidence but some uncertainty or conflicting research
  • Low Confidence (50-69%): Limited data, requires professional consultation
  • Insufficient Data (<50%): Recommendation deferred to human experts

Human Oversight and Review

Expert Review Process

🔍 Continuous Monitoring

All AI decisions monitored in real-time by our food safety team

  • High-risk alerts reviewed within 2 hours
  • Medium-risk recommendations reviewed daily
  • Low-risk decisions sampled weekly for quality assurance

👨‍⚕️ Expert Panel

Qualified professionals oversee AI decision-making:

  • Registered Nutritionists (UKVRN)
  • Food Safety Scientists
  • Paediatric Nutrition Specialists
  • Toxicology Consultants

⚡ Immediate Intervention

Human experts can immediately override AI decisions when:

  • New safety concerns emerge
  • Regulatory guidance changes
  • User reports adverse reactions
  • AI confidence levels are too low

Your Rights Regarding AI Decisions

UK GDPR Article 22 Rights

You have the right not to be subject to purely automated decision-making. For all AI-generated recommendations, you have the right to:

  • Human Review: Request human review of any AI recommendation
  • Explanation: Understand how and why the AI reached its decision
  • Challenge: Contest AI decisions you believe are incorrect
  • Alternative Assessment: Request a non-automated assessment from our experts
  • Opt-Out: Choose to receive only human-reviewed recommendations

How to Exercise Your Rights

📧 Email Request

Contact info@reclaimed.health with “AI Review Request” in the subject line

📱 In-App Feature

Use the “Request Human Review” button available on all AI recommendations

⚙️ Account Settings

Change your preferences to “Human Review Only” in your account settings

Response Times

  • Urgent Health Concerns: Within 2 hours during business hours
  • Routine Reviews: Within 1-2 business days
  • Complex Assessments: Within 5 business days
  • Account Setting Changes: Immediate effect

AI Ethics and Fairness

Ethical AI Principles

🎯 Purpose Limitation

AI used solely for food safety and health guidance, never for commercial profiling or discrimination

⚖️ Fairness and Non-Discrimination

Regular testing to ensure AI doesn’t discriminate based on age, ethnicity, socioeconomic status, or dietary culture

🔒 Privacy by Design

AI processing minimises personal data use and incorporates privacy protections from the ground up

🌍 Cultural Sensitivity

AI trained to respect diverse dietary practices and cultural food preferences across UK communities

Bias Prevention Measures

  • Diverse Training Data: Inclusion of foods from all major UK cultural communities
  • Regular Audits: Quarterly fairness testing across different demographic groups
  • Expert Diversity: Human oversight team represents diverse backgrounds and expertise
  • User Feedback: Continuous monitoring for reports of biased or inappropriate recommendations
  • Transparent Algorithms: Explainable AI techniques to identify and correct biases

Data Protection and AI

How AI Uses Your Data

Data Processing for AI:

  • Input Data: Food scans, ingredient lists, dietary preferences you provide
  • Processing: Anonymised analysis against safety databases
  • Learning: Aggregated, non-personal data used to improve AI accuracy
  • Storage: Personal AI interactions encrypted and stored securely
  • Retention: AI training data retained indefinitely, personal data as per our Privacy Policy

AI Data Protection Measures

  • Data Minimisation: AI uses only essential data needed for safety assessments
  • Pseudonymisation: Personal identifiers removed from AI training data
  • Secure Processing: All AI computations performed in secure, encrypted environments
  • Access Controls: Strict limits on who can access AI systems and data
  • Audit Trails: Complete logging of all AI decisions and data access

Continuous Improvement

How We Improve Our AI

📊 Performance Monitoring

Real-time tracking of AI accuracy, user satisfaction, and expert agreement rates

🔄 Regular Updates

Monthly model updates incorporating new research, regulatory changes, and user feedback

👥 User Research

Regular surveys and interviews with families to understand AI effectiveness and concerns

🎓 Expert Collaboration

Ongoing partnerships with UK universities and research institutions

Innovation Roadmap

  • Q1 2025: Enhanced allergen cross-reactivity detection
  • Q2 2025: Personalised nutrition recommendations for children
  • Q3 2025: Real-time regulatory alert integration
  • Q4 2025: Multilingual support for diverse UK communities
  • 2026: Predictive food safety trend analysis

Getting Help and Reporting Issues

AI Support and Questions

Contact Our AI Team:

  • Email: info@reclaimed.health
  • Subject Line: “AI Transparency Question” or “AI Review Request”
  • AI Ethics Officer: Ammar Jawad Doosh, CEO
  • Response Time: Within 2 business days for general questions, 2 hours for urgent safety concerns

Reporting AI Issues

Please report immediately if you experience:

  • Incorrect or potentially dangerous AI recommendations
  • Biased or discriminatory AI behaviour
  • AI decisions that seem to ignore your stated preferences
  • Technical issues with AI features
  • Concerns about AI transparency or fairness

Emergency Safety Concerns

If you believe our AI has provided dangerous health advice:

  • Immediate Action: Stop using the product and seek professional medical advice
  • Report Urgently: Email info@reclaimed.health with “URGENT AI SAFETY” in subject
  • Emergency Response: We investigate all safety reports within 2 hours
  • Follow-up: We will contact you within 24 hours with our investigation results

This AI Transparency Statement reflects our commitment to responsible AI development for food safety. We believe families deserve to understand exactly how AI is used to protect their health, and we’re committed to maintaining the highest standards of transparency, fairness, and safety in all our AI systems.

Document Version: 1.0

Review Date: January 3, 2026

Compliance: UK GDPR Article 22, EU AI Act, UK Algorithmic Transparency Standards

AI Ethics Officer: Ammar Jawad Doosh, CEO