Responsible AI Flashcards
6 cards from real Microsoft Azure AI Fundamentals practice questions. Tap to flip, then mark Knew It or Still Learning — missed cards come back until you master them.
Read the first 6 Responsible AI flashcards as text
Which Microsoft tool provides a dashboard for assessing error analysis, fairness, and model interpretability in Azure Machine Learning?
Answer: Responsible AI Dashboard in Azure Machine Learning
The Responsible AI Dashboard in Azure Machine Learning consolidates tools for error analysis, fairness assessment, model explanations, and data exploration in one interface.
Under Microsoft's Responsible AI principles, what should happen when an AI system causes unintended harm?
Answer: Accountability mechanisms should identify responsible parties, and remediation steps should be taken
The Accountability principle requires that organizations identify who is responsible when AI causes harm and establish clear processes for remediation and learning from failures.
What is 'data poisoning' in the context of AI security?
Answer: Intentionally introducing malicious or misleading data into a training dataset to corrupt the model's behavior
Data poisoning is an adversarial attack where malicious data is injected into a training set to manipulate the model's learned behavior in harmful ways.
Which concept describes the practice of testing AI systems with diverse and adversarial inputs before deployment?
Answer: Red teaming
Red teaming involves probing an AI system with adversarial, edge-case, or unexpected inputs to identify failure modes and vulnerabilities before deployment.
According to Microsoft's Responsible AI framework, who bears responsibility for ensuring AI systems are used ethically?
Answer: Both the organizations that build and deploy AI systems
The Accountability principle places responsibility on all organizations involved — both those that build AI systems and those that deploy them — to ensure ethical use.
What is the primary goal of 'differential privacy' in AI model training?
Answer: Protect individual data records from being inferred from a trained model's outputs
Differential privacy adds mathematical noise to training processes so that the model cannot reveal information about specific individuals in the training dataset.