CET-110 Exam Prep
CET-110 Exam Glossary - 35 Terms
Search the terminology pack for Certified Ethical Emerging Technologist. Use these definitions with the study guide and practice questions.
A
- Accountability
- The obligation of individuals or teams to answer for decisions, actions, and outcomes related to AI systems.
- AI lifecycle
- The full sequence of stages in AI development and use, from planning and data collection to deployment and monitoring.
- AI-driven content feed
- A content delivery system that uses AI to personalize, rank, or recommend information to users.
B
- Benchmarking
- Comparing current practices against standards or peers to assess performance and identify improvements.
- Bias mitigation
- The use of methods and controls to reduce unfair bias in data, models, and AI outcomes.
- Business imperatives
- Operational or commercial goals that drive organizational decision-making and product development.
- Business objective
- A clearly defined organizational goal that justifies a data collection or AI use activity.
D
- Data collection
- The process of gathering data for training, testing, or operating an AI system.
- Data minimization
- The practice of limiting data collection and processing to what is adequate, relevant, and necessary.
- Dataset linkage
- Combining a dataset with other data sources in ways that may reveal additional information or identities.
- De-identification
- The process of removing or altering identifiers in data to reduce the ability to link records to individuals.
- Decision rationale
- The recorded justification explaining why a particular governance or operational decision was made.
E
- Ethical AI governance
- The framework of policies, roles, processes, and oversight used to guide responsible AI development and use.
G
- Gap analysis
- The process of identifying differences between current practices and desired or expected standards.
H
- Harmful influence
- Negative impact on users' beliefs, decisions, or behavior caused by persuasive or manipulative system outputs.
- High-stakes AI system
- An AI system used in contexts where errors or bias can significantly affect people’s rights, opportunities, or wellbeing.
M
- Model design
- The selection and structuring of algorithms, features, and system architecture for an AI model.
N
- Necessity of collection
- The principle that personal data should be collected only when needed for a defined objective.
O
- Ongoing monitoring
- Continuous observation of an AI system after deployment to detect issues and maintain performance and compliance.
- Organizational maturity
- The level of development, consistency, and effectiveness of an organization’s AI ethics program or controls.
- Oversight functions
- Independent or supervisory roles that review, challenge, and monitor AI-related decisions and risks.
P
- Preprocessing
- The preparation and transformation of raw data before model training or analysis.
- Privacy by design
- An approach that embeds privacy safeguards into systems from project inception so compliance is addressed proactively.
- Privacy risk
- The possibility that personal data may be exposed, misused, or linked back to individuals.
R
- Re-identification
- The process by which de-identified data is linked with other information to infer or recover identity.
- Recognized standards
- Established external frameworks or guidelines used to evaluate and improve AI ethics practices.
- Regulatory compliance
- Conformance with laws, regulations, and formal requirements that govern business and AI practices.
- Residual privacy risk
- The remaining privacy risk that persists even after safeguards and reviews have been applied.
- Risk acceptance
- A documented decision to proceed despite remaining risk after evaluation and mitigation.
- Roles and responsibilities
- Clearly assigned duties and accountabilities across functions involved in AI oversight and operations.
S
- Safeguards
- Protective measures built into systems or processes to reduce harm, misuse, or unacceptable risk.
T
- Targeted improvement planning
- Prioritizing and organizing actions to address identified weaknesses in governance or ethics practices.
- Testing
- The evaluation of a system to assess performance, reliability, fairness, and other requirements before or after release.
- Transparency measures
- Methods used to make AI systems and their operation understandable to stakeholders.
U
- Use case
- A specific context or scenario in which a system is applied and evaluated.
About These Definitions
These definitions are loaded from the shared release pack. Use them with the study guide and practice questions to connect vocabulary to exam scenarios.