AI Glossary
Plain language definitions for AI terms used on MyAIConsent, aligned with ISO 22989 international standards.
AI System
An engineered system that generates outputs such as predictions, recommendations, decisions or content that influence real or virtual environments. On MyAIConsent, every AI Twin is an AI system.
AI Twin
A knowledge-based AI system trained exclusively on content provided and consented to by its creator. AI Twins on MyAIConsent use Retrieval Augmented Generation (RAG) to answer questions based solely on uploaded knowledge files.
Bias
Systematic unfairness in AI outputs that results from imbalanced training data or flawed model design. MyAIConsent mitigates bias by restricting AI Twin responses to creator-provided content only.
Consent
Freely given, specific, informed and unambiguous agreement by a person to the use of their content or personal data. MyAIConsent requires explicit consent before any content is used to train or inform an AI Twin.
Deployer
An organisation or individual that uses an AI system developed by another party. Under the EU AI Act, deployers carry obligations around appropriate use, monitoring and transparency.
Governance
The policies, processes and controls an organisation puts in place to ensure responsible development and use of AI systems. MyAIConsent aligns its governance framework with ISO 42001 (AI Management Systems) and ISO 42005 (AI Impact Assessment).
Hallucination
When an AI model generates plausible-sounding but factually incorrect information not grounded in its knowledge base. MyAIConsent reduces hallucination risk through RAG — responses cite actual uploaded documents.
Human Oversight
The ability of humans to monitor, intervene in and correct the behaviour of an AI system. On MyAIConsent, creators control all content, system guidelines and visibility settings for their AI Twins at all times.
Impact Assessment
A structured process for identifying and evaluating the potential effects of an AI system on individuals and society before deployment. ISO 42005 provides guidance on AI impact assessments. MyAIConsent requires an impact assessment for every AI Twin before it can be made public.
Provider
An organisation or individual that develops and places an AI system on the market. MyAIConsent is the provider of the AI Twin platform. Creators who publish AI Twins are deployers.
Provenance
The documented origin and chain of custody of data used in an AI system. Provenance tracking is central to MyAIConsent's long-term vision — every piece of content on the platform has a verified, consented origin.
RAG (Retrieval Augmented Generation)
A technique where an AI model retrieves relevant content from a knowledge base before generating a response. MyAIConsent uses RAG so that AI Twin responses are grounded in the creator's uploaded documents, not general internet knowledge.
Source Verification
The practice of showing which documents informed an AI response, along with a confidence score. MyAIConsent offers optional source verification on every AI Twin response.
Training Data
Data used to develop or fine-tune an AI model. On MyAIConsent, creator content is never used as training data for any AI model. It is used only for RAG retrieval within that creator's own AI Twin.
Transparency
The principle that the nature, capabilities and limitations of an AI system should be clearly communicated to users. MyAIConsent discloses the AI model used (Claude Sonnet 4 by Anthropic) and requires AI Twins to identify themselves as AI systems.
Definitions aligned with ISO 22989:2022 — Artificial intelligence — Concepts and terminology. Where platform-specific context applies, definitions have been adapted accordingly.