Accountability | The principle that ensures clear responsibility can be assigned for all decisions and actions throughout an AI system's lifecycle, from planning to decommissioning |
Affinity bias | When someone prefers individuals who are similar to them in terms of ideology, attitudes, appearance, or religion |
Agile | A project management and development approach that emphasizes flexibility, iterative progress, and collaboration |
AI governance | The framework of policies, structures, and processes created to maximize the benefits of AI while minimizing its risks |
AI literacy | The ability to understand, critically evaluate, and effectively interact with AI technologies, including knowledge of AI's capabilities, limitations, and potential impacts |
AI PPM (AI Project Portfolio Management) | The centralized management of an organization's AI initiatives to meet strategic objectives by optimizing resource allocation, balancing risks, and maximizing value |
Algorithm | A set of rules or instructions given to an AI system to help it learn, make decisions, and solve problems |
Amplification bias | Occurs when several AI systems, each with separate biases, interact and mutually reinforce each other's biases |
API (Application Programming Interface) | A set of rules and protocols that allows different software applications to communicate with each other |
Automation bias | When conclusions drawn from algorithms are valued more highly than human analyses |
Cloud storage | The practice of storing data and applications on remote servers accessed via the internet, rather than on local computers |
Coverage bias | A form of sampling bias that occurs when a selected population does not match the intended population |
Data architecture | The overall structure of an organization's data assets and data management resources |
Data bias | A type of error where certain elements of a data set are more heavily weighted or represented than others |
Data cleaning | The process of detecting and correcting errors, inconsistencies, and inaccuracies in data sets |
Data governance | The framework of policies, processes, and standards that ensure the effective management of data assets |
Data literacy | The ability to read, understand, create, and communicate data as information, including understanding data collection, analysis, interpretation, and presentation |
Data migration | The process of moving data from one system or storage type to another |
Data mining | The process of discovering patterns and relationships in large data sets |
Data set | A collection of related data points or information used to train AI systems |
Data steward | A person responsible for managing and overseeing an organization's data assets |
Data visualization | The graphical representation of data and information using charts, graphs, and other visual elements |
Database | An organized collection of structured information or data |
Deepfake | Synthetic media where a person's likeness is replaced with someone else's using AI |
Deployment | The process of making an AI system or application available for use |
Deployment bias | When a system that works well in a test environment performs poorly when deployed in the real world |
Ethical principles | Guidelines that ensure AI systems respect privacy, transparency, accountability, fairness, and human autonomy while promoting societal well-being |
Explainability | The ability for humans to understand and trust decisions, recommendations, or predictions made by an AI systems |
Feedback loop bias | When the output of an AI system influences future inputs, potentially reinforcing existing biases |
Generative AI | AI systems capable of creating new content (text, images, code, etc.) based on patterns learned from training data |
HIC (Human-in-Command) | A comprehensive oversight approach that considers broader economic, social, legal, and ethical impacts of AI systems |
HITL (Human-in-the-Loop) | An approach where a human mediates all decisions made by the AI system |
HOTL (Human-on-the-Loop) | An approach where humans monitor AI system operations and can intervene when necessary |
Infrastructure | The hardware, software, networks, and facilities that support an organization's IT operations |
Intellectual property rights | Rights that protect the investment of rights-holders in original content, including copyrights and accessory rights |
KPI (Key Performance Indicator) | Measurable values that demonstrate how effectively an organization is achieving key objectives |
LLM (Large Language Model) | AI models trained on vast amounts of text data that can understand and generate human-like text |
Linguistic bias | When an AI algorithm favours certain linguistic styles, vocabularies, or cultural references over others |
Machine learning | A subset of AI that enables systems to learn and improve from experience without explicit programming |
Natural Language Processing (NLP) | The ability of computers to understand, interpret, and generate human language |
Neural network | A computer system modelled on the human brain, designed to recognize patterns |
Open source | Software whose source code is freely available for anyone to inspect, modify, and enhance |
Participation bias | A form of sampling bias that occurs when certain groups choose not to participate in data collection |
Pilot project | A small-scale preliminary study to evaluate feasibility, cost, and potential issues |
Privacy | The principle that AI systems should respect and protect personal data and information |
Prompt engineering | The practice of designing and optimizing inputs to AI systems to generate desired outputs |
Proxy bias | When variables used as proxies for protected attributes introduce bias into the model |
RPA (Robotic Process Automation) | Technology that automates repetitive tasks through software robots |
Robustness | The ability of AI systems to maintain reliable operation and withstand adverse conditions or attacks |
Sampling bias | When data is not randomly selected, resulting in a sample that is not representative of the population |
Shadow AI | The unsupervised or unsanctioned use of AI tools within an organization outside of its IT and cybersecurity framework |
Shadow IT | The use of IT systems, devices, software, or services without explicit organizational approval |
Stakeholder | Any person, group, or organization that has an interest in or is affected by an AI project |
SVG (Scalable Vector Graphics) | A web-friendly vector image format that can scale without losing quality |
Temporal bias | When training data becomes outdated and no longer represents current realities |
Traceability | The ability to follow and monitor the entire lifecycle of an AI system |
Training data | The initial data set used to teach an AI system to perform its intended function |
Transparency | The communication of appropriate information about AI systems in an understandable and accessible format |
Use case | A specific situation or scenario where an AI system or application could be used |
XAI (eXplainable AI) | AI systems designed to be interpretable and understandable by humans |
