[{"data":1,"prerenderedAt":66},["ShallowReactive",2],{"policy-ai-use-transparency-policy":3},{"html":4,"meta":5,"toc":24,"summary":64,"error":65},"\n\u003Ch2 id=\"1-purpose-and-scope\">1. Purpose and Scope\u003C\u002Fh2>\n\u003Cp>This policy establishes how Qualitect Ltd uses artificial intelligence technologies within the Qualitect platform. It sets out our commitment to transparent, responsible and accountable AI use in the design and development of qualifications. This policy applies to all users of the Qualitect platform and covers interactions with third-party AI service providers used by the platform.\u003C\u002Fp>\n\u003Ch2 id=\"2-definitions\">2. Definitions\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>AI-\u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"generated-content\" data-glossary-term=\"Generated content\" aria-label=\"Definition: Generated content\">Generated Content\u003C\u002Fbutton>\u003C\u002Fstrong> means any text, data or recommendations produced by artificial intelligence models deployed within the Qualitect platform.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Evidence Retrieval\u003C\u002Fstrong> means the automated process of identifying and ranking relevant evidence using \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"vector-embeddings\" data-glossary-term=\"Vector embeddings\" aria-label=\"Definition: Vector embeddings\">vector embeddings\u003C\u002Fbutton> and machine learning techniques.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>\u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"generation-pipeline\" data-glossary-term=\"Generation pipeline\" aria-label=\"Definition: Generation pipeline\">Generation Pipeline\u003C\u002Fbutton>\u003C\u002Fstrong> means the end-to-end process that combines AI models, prompt engineering, evidence retrieval and compliance validation to produce \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"qualification\" data-glossary-term=\"Qualification\" aria-label=\"Definition: Qualification\">qualification\u003C\u002Fbutton> content.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Human-in-the-Loop\u003C\u002Fstrong> means the mandatory requirement for \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"human-review\" data-glossary-term=\"Human review\" aria-label=\"Definition: Human review\">human review\u003C\u002Fbutton> and explicit acceptance of all AI-generated content before it can be saved or submitted.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>\u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"prompt-hash\" data-glossary-term=\"Prompt hash\" aria-label=\"Definition: Prompt hash\">Prompt Hash\u003C\u002Fbutton>\u003C\u002Fstrong> means a cryptographic fingerprint of the instructions sent to AI models, used for \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"audit-trail\" data-glossary-term=\"Audit trail\" aria-label=\"Definition: Audit trail\">audit trail\u003C\u002Fbutton> purposes without storing the full prompt content.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>\u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"syllabus\" data-glossary-term=\"Syllabus\" aria-label=\"Definition: Syllabus\">Syllabus\u003C\u002Fbutton>\u003C\u002Fstrong> means a document that sets out the content, structure and requirements of a qualification. Sometimes also known as a specification or curriculum.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Qualification\u003C\u002Fstrong> means a formal recognition that a \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"learner\" data-glossary-term=\"Learner\" aria-label=\"Definition: Learner\">learner\u003C\u002Fbutton> has achieved a defined standard of knowledge, skill or competence.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>\u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"rag-pipeline\" data-glossary-term=\"RAG Pipeline\" aria-label=\"Definition: RAG Pipeline\">RAG Pipeline\u003C\u002Fbutton>\u003C\u002Fstrong> means Retrieval-Augmented Generation, the process that combines relevant evidence documents with AI generation to produce contextually informed qualification content.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Vector Embeddings\u003C\u002Fstrong> means numerical representations of text that capture meaning and context, enabling the platform to find content that is meaningfully similar rather than relying on keyword matching alone.\u003C\u002Fp>\n\u003Ch2 id=\"3-ai-technologies-used\">3. AI Technologies Used\u003C\u002Fh2>\n\u003Ch3 id=\"3-1-primary-ai-systems\">3.1 Primary AI Systems\u003C\u002Fh3>\n\u003Cp>Qualification generation uses large language models to produce draft \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"learning-outcomes\" data-glossary-term=\"Learning outcomes\" aria-label=\"Definition: Learning outcomes\">learning outcomes\u003C\u002Fbutton>, \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"assessment-criteria\" data-glossary-term=\"Assessment criteria\" aria-label=\"Definition: Assessment criteria\">assessment criteria\u003C\u002Fbutton> and \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"indicative-content\" data-glossary-term=\"Indicative content\" aria-label=\"Definition: Indicative content\">indicative content\u003C\u002Fbutton>. Evidence retrieval uses vector embeddings to identify relevant documents. Compliance validation support combines AI with deterministic rules engines. An AI assistant feature provides contextual help.\u003C\u002Fp>\n\u003Ch3 id=\"3-2-ai-model-providers\">3.2 AI Model Providers\u003C\u002Fh3>\n\u003Cp>\u003Cstrong>Primary Provider:\u003C\u002Fstrong> Anthropic Claude serves as our primary large language model through the Anthropic \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"api\" data-glossary-term=\"API\" aria-label=\"Definition: API\">API\u003C\u002Fbutton>.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Fallback Provider:\u003C\u002Fstrong> OpenAI serves as our fallback large language model provider, engaged automatically when the primary provider experiences availability issues.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Embedding Provider:\u003C\u002Fstrong> OpenAI's \u003Ccode>text-embedding-3-small\u003C\u002Fcode> model is used to generate vector embeddings of uploaded evidence files. Embedding generation is a one-shot transform per chunk and produces numerical representations that cannot be reversed to the original text.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Reranker Provider:\u003C\u002Fstrong> Cohere Rerank is used to improve the relevance ordering of retrieved evidence chunks before they are passed to the primary or fallback LLM. The reranker operates on short text snippets only and does not retain the inputs.\u003C\u002Fp>\n\u003Cp>Our AI providers do not use Qualitect data to train their models. Data processed for generation is retained by providers only for a limited period for abuse monitoring under each provider's terms and is then deleted, except where retention is required by law. The same oversight and validation requirements apply to all AI providers used by the platform. Each provider is engaged under a \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"data-processing-agreement\" data-glossary-term=\"Data Processing Agreement\" aria-label=\"Definition: Data Processing Agreement\">Data Processing Agreement\u003C\u002Fbutton>.\u003C\u002Fp>\n\u003Ch3 id=\"3-3-technical-architecture\">3.3 Technical Architecture\u003C\u002Fh3>\n\u003Cp>All AI processing occurs through encrypted connections using \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"tls\" data-glossary-term=\"TLS\" aria-label=\"Definition: TLS\">TLS\u003C\u002Fbutton> 1.3. The platform maintains an append-only audit trail of every AI interaction including model selection, generation parameters and prompt hashes.\u003C\u002Fp>\n\u003Ch2 id=\"4-human-oversight-and-control\">4. Human Oversight and Control\u003C\u002Fh2>\n\u003Ch3 id=\"4-1-mandatory-human-review\">4.1 Mandatory Human Review\u003C\u002Fh3>\n\u003Cp>Every piece of AI-generated content must undergo human review before it becomes part of a qualification. The platform enforces this at the API level, preventing automated acceptance. Users must explicitly accept or reject each generated element.\u003C\u002Fp>\n\u003Ch3 id=\"4-2-override-and-escalation\">4.2 Override and Escalation\u003C\u002Fh3>\n\u003Cp>Users can override compliance validation findings with documented justification. All override decisions are recorded in the audit trail. Where Platform Users identify AI-generated content that may pose compliance risks, they should report these through the Complaints Procedure.\u003C\u002Fp>\n\u003Ch3 id=\"4-3-quality-monitoring\">4.3 Quality Monitoring\u003C\u002Fh3>\n\u003Cp>The platform continuously monitors quality of AI-generated content. Bias detection operates through multiple mechanisms including \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"verb-level-matrix\" data-glossary-term=\"Verb-level matrix\" aria-label=\"Definition: Verb-level matrix\">verb-level matrix\u003C\u002Fbutton> enforcement.\u003C\u002Fp>\n\u003Ch2 id=\"5-data-handling-in-ai-processing\">5. Data Handling in AI Processing\u003C\u002Fh2>\n\u003Ch3 id=\"5-1-data-sent-to-ai-providers\">5.1 Data Sent to AI Providers\u003C\u002Fh3>\n\u003Cp>The data sent to AI providers is limited to qualification parameters, prompts and extracts from uploaded evidence. Platform Users should minimise the \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"personal-data\" data-glossary-term=\"Personal data\" aria-label=\"Definition: Personal data\">personal data\u003C\u002Fbutton> included in uploaded material; where personal data is included, it is processed as described in this policy and, for customer organisations, the Data Processing Agreement.\u003C\u002Fp>\n\u003Ch3 id=\"5-2-data-not-sent-to-ai-providers\">5.2 Data Not Sent to AI Providers\u003C\u002Fh3>\n\u003Cp>User personal data, confidential commercial information and complete evidence files are not sent. Learner data is never transmitted under any circumstances.\u003C\u002Fp>\n\u003Ch3 id=\"5-3-data-processing-agreements\">5.3 Data Processing Agreements\u003C\u002Fh3>\n\u003Cp>Qualitect maintains comprehensive data processing agreements with all AI service providers specifying data handling requirements and security obligations.\u003C\u002Fp>\n\u003Ch2 id=\"6-transparency-measures\">6. Transparency Measures\u003C\u002Fh2>\n\u003Ch3 id=\"6-1-content-labelling\">6.1 Content Labelling\u003C\u002Fh3>\n\u003Cp>All AI-generated content is clearly labelled. Generation indicators show the generating model, timestamp and evidence sources.\u003C\u002Fp>\n\u003Ch3 id=\"6-2-audit-trail-transparency\">6.2 Audit Trail Transparency\u003C\u002Fh3>\n\u003Cp>Comprehensive audit trails of all AI interactions are maintained. Prompt hash recording enables audit capabilities without storing sensitive prompt content.\u003C\u002Fp>\n\u003Ch3 id=\"6-3-explainability\">6.3 Explainability\u003C\u002Fh3>\n\u003Cp>Users can access detailed explanations of how AI-generated content relates to the evidence base.\u003C\u002Fp>\n\u003Ch2 id=\"7-bias-prevention-and-monitoring\">7. Bias Prevention and Monitoring\u003C\u002Fh2>\n\u003Ch3 id=\"7-1-structural-bias-controls\">7.1 Structural Bias Controls\u003C\u002Fh3>\n\u003Cp>Verb-level matrix enforcement ensures appropriate cognitive demand. Subject benchmark alignment requires content to align with established frameworks.\u003C\u002Fp>\n\u003Ch3 id=\"7-2-content-review-for-bias\">7.2 Content Review for Bias\u003C\u002Fh3>\n\u003Cp>Human reviewers must consider potential bias including demographic assumptions, cultural bias and accessibility barriers.\u003C\u002Fp>\n\u003Ch3 id=\"7-3-ongoing-monitoring\">7.3 Ongoing Monitoring\u003C\u002Fh3>\n\u003Cp>Regular analysis of generated content identifies potential bias patterns. Users can report concerns through feedback mechanisms.\u003C\u002Fp>\n\u003Ch2 id=\"8-limitations-and-risks\">8. Limitations and Risks\u003C\u002Fh2>\n\u003Ch3 id=\"8-1-content-accuracy\">8.1 Content Accuracy\u003C\u002Fh3>\n\u003Cp>AI-generated content may contain errors despite compliance validation. Users must exercise professional judgement in reviewing and accepting generated content.\u003C\u002Fp>\n\u003Ch3 id=\"8-2-compliance-support\">8.2 Compliance Support\u003C\u002Fh3>\n\u003Cp>The platform provides tools to support compliance but does not guarantee regulatory approval. Generated content requires thorough human review before submission to any relevant regulatory body.\u003C\u002Fp>\n\u003Ch3 id=\"8-3-technical-risk-management\">8.3 Technical Risk Management\u003C\u002Fh3>\n\u003Cp>Fallback systems, error handling and monitoring capabilities manage technical risks.\u003C\u002Fp>\n\u003Ch2 id=\"9-accountability\">9. Accountability\u003C\u002Fh2>\n\u003Ch3 id=\"9-1-qualitect-responsibilities\">9.1 Qualitect Responsibilities\u003C\u002Fh3>\n\u003Cp>Qualitect accepts responsibility for the design, implementation and operation of AI systems within the platform.\u003C\u002Fp>\n\u003Ch3 id=\"9-2-user-responsibilities\">9.2 User Responsibilities\u003C\u002Fh3>\n\u003Cp>Users retain responsibility for qualification content regardless of whether AI assistance was used. Professional expertise in reviewing AI-generated content remains essential.\u003C\u002Fp>\n\u003Ch2 id=\"10-alignment-with-uk-ai-principles\">10. Alignment with UK AI Principles\u003C\u002Fh2>\n\u003Ch3 id=\"10-1-safety-security-and-robustness\">10.1 Safety, Security and Robustness\u003C\u002Fh3>\n\u003Cp>Comprehensive security measures including encryption, access controls and monitoring.\u003C\u002Fp>\n\u003Ch3 id=\"10-2-transparency-and-explainability\">10.2 Transparency and Explainability\u003C\u002Fh3>\n\u003Cp>Clear information about how AI systems operate, what data is processed and how content relates to evidence sources.\u003C\u002Fp>\n\u003Ch3 id=\"10-3-fairness-and-non-discrimination\">10.3 Fairness and Non-Discrimination\u003C\u002Fh3>\n\u003Cp>Multiple measures to prevent bias through technical controls, human review and ongoing monitoring.\u003C\u002Fp>\n\u003Ch3 id=\"10-4-accountability-and-governance\">10.4 Accountability and Governance\u003C\u002Fh3>\n\u003Cp>Clear accountability frameworks for AI system operation and content quality. Our AI governance framework is designed to align with the requirements of ISO 42001, the international standard for AI management systems and with the broader requirements of ISO 9001 and ISO 21001 as they apply to the quality and educational integrity of AI-generated qualification content. Qualitect is working towards formal accreditation against ISO standards relevant to its operations. References to ISO alignment describe the framework adopted and do not indicate current certification.\u003C\u002Fp>\n\u003Ch3 id=\"10-5-contestability-and-redress\">10.5 Contestability and Redress\u003C\u002Fh3>\n\u003Cp>Users can challenge AI-generated content and system decisions through the Complaints Procedure.\u003C\u002Fp>\n\u003Ch2 id=\"11-roles-and-responsibilities\">11. Roles and Responsibilities\u003C\u002Fh2>\n\u003Cp>Platform Users are responsible for reviewing and explicitly accepting all AI-generated content and applying professional judgement. Platform Users with \u003Cbutton type=\"button\" class=\"q-glossary-trigger\" data-glossary-slug=\"quality-assurance\" data-glossary-term=\"Quality assurance\" aria-label=\"Definition: Quality assurance\">quality assurance\u003C\u002Fbutton> obligations must include AI-generated content within their quality assurance processes. The Compliance and Accreditation Lead has day-to-day responsibility for this policy. The Development Team is responsible for maintaining the technical AI infrastructure.\u003C\u002Fp>\n\u003Ch2 id=\"12-related-policies\">12. Related Policies\u003C\u002Fh2>\n\u003Cp>Privacy Policy, Data Protection Policy, Data and Content Sourcing Policy, Quality Assurance Policy, Complaints Procedure.\u003C\u002Fp>\n\u003Ch2 id=\"13-contact-information\">13. Contact Information\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>General:\u003C\u002Fstrong> hello@qualitect.co.uk\u003Cbr>\u003Cstrong>Data protection contact:\u003C\u002Fstrong> dataprotection@qualitect.co.uk\u003Cbr>\u003Cstrong>Complaints:\u003C\u002Fstrong> complaints@qualitect.co.uk\u003Cbr>\u003Cstrong>Registered office:\u003C\u002Fstrong> Qualitect Limited, 32 Willoughby Road, London, N8 0JG\u003Cbr>\u003Cstrong>Trading address:\u003C\u002Fstrong> 8a Stafford Street, London, W1S 4RU\u003Cbr>\u003Cstrong>Website:\u003C\u002Fstrong> qualitect.co.uk\u003C\u002Fp>",[6,9,12,15,18,21],{"k":7,"v":8},"Version","2.0",{"k":10,"v":11},"Status","Live",{"k":13,"v":14},"Effective Date","13 July 2026",{"k":16,"v":17},"Review Date","13 July 2027",{"k":19,"v":20},"Policy Owner","Compliance and Accreditation Lead",{"k":22,"v":23},"Approved By","Chief Executive Officer",[25,28,31,34,37,40,43,46,49,52,55,58,61],{"id":26,"text":27},"1-purpose-and-scope","1. Purpose and Scope",{"id":29,"text":30},"2-definitions","2. Definitions",{"id":32,"text":33},"3-ai-technologies-used","3. AI Technologies Used",{"id":35,"text":36},"4-human-oversight-and-control","4. Human Oversight and Control",{"id":38,"text":39},"5-data-handling-in-ai-processing","5. Data Handling in AI Processing",{"id":41,"text":42},"6-transparency-measures","6. Transparency Measures",{"id":44,"text":45},"7-bias-prevention-and-monitoring","7. Bias Prevention and Monitoring",{"id":47,"text":48},"8-limitations-and-risks","8. Limitations and Risks",{"id":50,"text":51},"9-accountability","9. Accountability",{"id":53,"text":54},"10-alignment-with-uk-ai-principles","10. Alignment with UK AI Principles",{"id":56,"text":57},"11-roles-and-responsibilities","11. Roles and Responsibilities",{"id":59,"text":60},"12-related-policies","12. Related Policies",{"id":62,"text":63},"13-contact-information","13. Contact Information","This policy establishes how Qualitect Ltd uses artificial intelligence technologies within the Qualitect platform. It sets out our commitment to trans",false,1784287572772]