Glossary
AI governance glossary
Core vocabulary for AI governance, policy and public-interest practice. Each term has a dedicated page, related entries, and links to authoritative sources. Disputed language is not treated as settled doctrine.
40 entries
- AI GovernanceThe institutions, rules, processes and practices used to steer the design, deployment and use of artificial intelligence toward public-interest outcomes.
- AI RegulationBinding or proposed legal rules that constrain the development, placing on the market, or use of AI systems.
- AI EthicsNormative inquiry into what constitutes acceptable design and use of AI, including values such as dignity, justice, autonomy and the public good.
- Responsible AIA practice-oriented label for designing, deploying and overseeing AI systems in ways intended to reduce harm and align with stated values, rights and laws.
- Trustworthy AIA policy term, especially in European documents, for AI that is lawful, ethical and technically robust.
- AI SafetyWork aimed at preventing AI systems from causing severe harm, including accidents, misuse and loss of meaningful human control.
- AI AssuranceIndependent or internal evidence-gathering that an AI system meets specified claims, standards or legal requirements.
- Algorithmic AccountabilityThe expectation that organizations can explain, justify and remedy the effects of algorithmic systems they use.
- Algorithmic BiasSystematic skew in an algorithmic system’s outputs that can produce unjustified disparate effects across people or groups.
- Algorithmic Impact AssessmentA structured review of an algorithmic system’s intended use, risks and affected communities, often required before public-sector deployment.
- AI Impact AssessmentAn evaluation of how an AI system may affect rights, safety, equality, the environment or public interests over its lifecycle.
- AI AuditA systematic examination of an AI system, its data, governance or outcomes against defined criteria.
- AI TransparencyThe availability of meaningful information about an AI system’s existence, purpose, operation, limitations and governance.
- AI ExplainabilityThe ability to provide reasons, in a form suitable for a given audience, for how an AI system arrived at an output.
- Human OversightArrangements that keep people able to understand, intervene in, or override an AI system’s operation where required.
- Human-Centred AIAn approach that treats human rights, human agency and social context as design constraints rather than afterthoughts.
- Risk-Based RegulationA regulatory design that scales legal duties according to estimated severity and likelihood of harm.
- High-Risk AIA legal category, notably in EU law, for AI systems whose use in specified contexts triggers enhanced obligations.
- Foundation ModelA large machine-learning model trained on broad data that can be adapted to many downstream tasks.
- General-Purpose AIAI models or systems designed to perform a wide range of distinct tasks, rather than a single predefined use.
- Generative AIAI systems that produce text, images, audio, code or other content that resembles material they were trained on.
- AI LiteracyThe knowledge and skills people need to understand, question and use AI systems in their roles as workers, citizens and consumers.
- AI StandardsDocumented technical or management specifications developed by standards bodies or consortia to support interoperability, quality and risk management.
- AI ProcurementThe rules and practices public and private buyers use to acquire AI systems, data or related services.
- Data GovernanceThe policies, rights and infrastructures that determine how data is collected, accessed, shared, protected and deleted.
- AI and Human RightsThe application of international human-rights law and standards to the design, deployment and governance of AI.
- AI and DemocracyThe effects of AI on elections, public debate, political equality, civic institutions and the conditions of self-government.
- AI and ElectionsThe use of AI in electoral processes and political campaigning, including risks to integrity, equality of political speech and voter privacy.
- AI and Civil SocietyThe roles of non-governmental organizations, movements and independent media in contesting, shaping and using AI.
- Digital RightsHuman rights as exercised in digital environments, including privacy, expression, association, equality and access to information.
- AI SovereigntyA political claim that a state or community should control critical AI infrastructure, data and policy choices rather than depend entirely on external providers.
- AI Capacity BuildingEfforts to strengthen the skills, institutions and infrastructure needed for societies to govern and benefit from AI.
- AI Incident ReportingThe practice of documenting and sharing information about AI system failures, misuse or harms so that others can learn and respond.
- AI Red-TeamingAdversarial testing that tries to elicit harmful, insecure or policy-violating behavior from an AI system.
- Model EvaluationSystematic measurement of an AI model’s capabilities, limitations, risks and performance on defined tasks.
- AI Governance FrameworkA structured set of principles, processes or controls intended to guide AI oversight inside an organization or across institutions.
- AI Safety GovernanceInstitutional arrangements aimed specifically at identifying, reducing and overseeing severe AI-related risks.
- Open AI ModelsAI models released with some combination of weights, code, data or documentation that others can inspect, adapt or deploy.
- AI and Public PolicyThe use of public authority—law, budget, institutions and diplomacy—to shape how AI is developed and used in society.
- AI in the Global SouthThe distinct opportunities, dependencies and governance challenges of AI in countries and communities historically marginalized in global technology rule-making.
Have an official resource, framework, law or other source to propose for a section of this directory? Use the suggestion form (opens in a new tab). Inclusion is editorial and not guaranteed.