All terms

• accountability
• accuracy
• adversarial example
• aggregation
• ai literacy
• ai office
• ai principles
• ai regulatory sandbox
• ai system
• algorithm
• alignment
• anonymization
• artificial intelligence
• artificial intelligence (ai) system
• artificial narrow intelligence (ani)
• attack
• audit
• authorised representative
• automated decision-making
• automation
• availability
• benchmark
• bias
• big data
• binding corporate rules
• biometric categorisation system
• biometric data
• biometric identification
• biometric verification
• ce marking
• common specification
• compliance
• confidentiality
• conformity assessment
• conformity assessment body
• consent
• constraint
• context
• control class
• controller
• correlation
• critical infrastructure
• cross-border processing
• data
• data analytics
• data concerning health
• data governance
• data literacy
• data minimization
• data provenance
• data quality
• data science
• data scientist
• data sharing
• dataset
• deep fake
• deep learning
• deployer
• deterministic
• discrimination
• distributor
• downstream provider
• drift
• emotion recognition system
• enterprise
• error
• experiment
• expert system
• explainability
• explanation
• exploratory
• facial recognition (fr)
• fairness
• filing system
• fine-tuning
• fitting
• floating-point operation
• foundation model
• general-purpose ai model
• general-purpose ai system
• generalization
• generative AI
• generative adversarial network (gan)
• genetic data
• global
• ground truth
• group of undertakings
• hallucination
• harmonised standard
• high-impact capabilities
• human oversight
• human-cognitive bias
• human-in-the-loop
• hypothesis testing
• impact assessment
• importer
• independence
• inference
• information society service
• informed consent
• input data
• instructions for use
• intended purpose
• international organisation
• interoperability
• interpretability
• large language model
• law enforcement
• local
• machine learning
• making available on the market
• market surveillance authority
• materiality
• metadata
• missing data
• model
• national competent authority
• non-personal data
• normalization
• notified body
• notifying authority
• open data
• operator
• outlier
• overfitting
• parity
• performance of an ai system
• personal data
• personal data breach
• placing on the market
• post-hoc explanation
• post-market monitoring system
• post-remote biometric identification system
• privacy
• processing
• processor
• profiling
• prompt
• prototype
• provider
• pseudo-anonymization (pseudonymization)
• pseudonymisation
• pseudonymization
• publicly accessible space
• putting into service
• quality
• real-time remote biometric identification system
• real-world testing plan
• reasonably foreseeable misuse
• recall of an ai system
• recipient
• relevant and reasoned objection
• reliability
• remote biometric identification system
• representative
• reproducibility
• resilience
• restriction of processing
• risk
• risk tolerance
• robustness
• safety component
• sampling
• sandbox plan
• sensitive operational data
• sensitivity analysis
• serious incident
• socio-technical system
• special categories of personal data
• standard deviation
• statistics
• stochastic
• structured data
• substantial modification
• supervisory authority
• supervisory authority concerned
• surveillance
• synthetic data
• systemic risk
• taxonomy
• technical control
• testing data
• testing in real-world conditions
• third party
• training data
• transparency
• trust
• trustworthiness
• trustworthy ai
• uncertainty
• validation
• validation data
• validation data set
• variable
• variance
• verifiable
• widespread infringement
• withdrawal of an ai system