Free Certification Practice Questions

CSA-TAISE

Loading…
Which of the following is a characteristic of Generative Adversarial Networks (GANs) relevant to AI safety?
#31
In the NIST AI RMF, which activity involves prioritizing and addressing identified AI risks through risk treatment options such as avoidance, mitigation, transfer, or acceptance?
#32
Which of the following is the MOST appropriate control to detect when a deployed AI model begins to degrade in performance due to data drift?
#33
A company builds an AI-powered recruitment tool that interviews candidates via video and scores them on personality traits inferred from facial expressions. Under the EU AI Act, this system would most likely be classified as:
#34
Which of the following is an example of AI being applied to cybersecurity offense rather than defense?
#35
Zero Trust applied to AI workloads means that:
#36
ISO/IEC 42001 is best described as:
#37
Which of the following describes a 'shadow AI' risk in an enterprise?
#38
Which of the following is the MOST important first step when developing an AI Impact Assessment (AIIA) for a new AI project?
#39
In the context of AI security, 'model poisoning' specifically refers to:
#40
Training data anonymization through k-anonymity ensures that:
#41
Which of the following is a key characteristic of Reinforcement Learning from Human Feedback (RLHF) used to align LLMs?
#42
The TAISE certification exam consists of how many questions, what is the time limit, and what is the passing score?
#43
When evaluating a vendor's AI product for enterprise procurement, which of the following is MOST important from an AI governance perspective?
#44
In the context of AI, what does the term 'hallucination' refer to?
#45
Which NIST AI RMF function is responsible for establishing organizational policies, roles, and accountability structures for AI risk?
#46
What does the 'right to erasure' (right to be forgotten) mean for organizations that have trained AI models on personal data?
#47
Which technique is used to explain the contribution of each input feature to a model's prediction by computing Shapley values from cooperative game theory?
#48
Output filtering and content moderation guardrails for LLMs are best applied:
#49
An AI system that penalizes loan applicants from certain ZIP codes—even though race is not a direct input—may exhibit:
#50
The practice of testing an AI system with a diverse set of real-world user inputs before public release is called:
#51
Which of the following best describes 'model collapse' in the context of training future AI models on AI-generated content?
#52
LIME (Local Interpretable Model-agnostic Explanations) is used to:
#53
The 'context window' of a large language model refers to:
#54
Jailbreaking an LLM refers to:
#55
Which transformer component is most directly responsible for capturing relationships between all tokens in an input sequence simultaneously?
#56
Which defense technique adds carefully calibrated random noise to training data or model outputs to protect individual privacy?
#57
The NIST AI RMF Generative AI Profile (NIST AI 600-1) introduces which AI-specific risk not covered in the original AI RMF 1.0?
#58
Which of the following best describes 'model cards' in the context of AI transparency?
#59
Which of the following describes the key distinction between 'AI safety' focused on long-term existential risks versus near-term AI safety?
#60