AI Transparency Vocabulary
Practice AI transparency vocabulary: explainability reports, black box models, model documentation, flagging decisions for human review, and explaining model decisions.
0 / 25 completed
1 / 25
The compliance team requests an ___ report for the loan decision model. What does this document contain?
An explainability report describes how a model makes decisions — which input features matter most, how they influence outputs, and how users or regulators can understand individual predictions. Required for high-risk AI systems.
2 / 25
A regulator calls your model a '___ box model.' What are they criticising?
A black box model is one whose internal workings are not interpretable — you can see the inputs and outputs but not the reasoning. Deep neural networks are often described this way, raising concerns for regulated domains.
3 / 25
The team publishes ___ documentation covering the model's training data, intended use, and limitations.
Model documentation (often a model card) is a standardised document that describes what a model does, what data it was trained on, its performance across subgroups, known limitations, and intended and prohibited uses.
4 / 25
The system logs show: 'The decision was ___ for human review.' What happened?
When a decision is flagged for human review, the AI system has identified a case that exceeds a confidence threshold or matches a sensitivity rule, routing it to a human reviewer instead of acting automatically.
5 / 25
In a stakeholder meeting you say: 'We can ___ 85% of decisions.' What are you communicating about the model?
'We can explain 85% of decisions' means that for 85% of model outputs, the system can provide an interpretable reason — such as the top contributing features — while the remaining 15% are still effectively a black box.
6 / 25
Reviewer: 'I'm seeing a high degree of disagreement between the model's predictions and the actual outcomes in this fraud detection cohort. The explanations are also quite opaque – it's hard to understand *why* the model flagged these transactions as suspicious. PR Description: 'This PR implements the new fraud detection model, leveraging a gradient boosting algorithm for improved accuracy.'Which of the following best describes the reviewer's concern regarding this PR?
The reviewer is highlighting a lack of transparency, specifically referring to 'opaque' decision-making. This term describes models where the reasoning behind predictions isn't clear or easily understood – often due to complex algorithms like gradient boosting. The options presented focus on different potential issues (overfitting, bias, insufficient documentation), but only 'opaque' directly addresses the core concern of difficult-to-interpret model behavior. It's crucial for developers to be able to articulate this lack of transparency when discussing models with stakeholders and reviewers.
7 / 25
Reviewer: 'The model's output is consistently exhibiting bias across different demographic groups. While the technical documentation describes the training data preprocessing steps, it lacks a clear articulation of how potential biases were identified and mitigated – essentially, there's no transparency into the fairness evaluation process. PR Description: 'This PR updates the API endpoint to include confidence scores for each prediction.' Which phrase best captures the reviewer's primary concern?The reviewer isn't primarily focused on the technical details of the API documentation (option A) or the accuracy of the confidence scores (option C). Instead, they are highlighting a critical lack of transparency regarding bias mitigation. The core issue is the absence of documented strategies for identifying and addressing potential biases within the model's training data and outputs – this directly relates to fairness concerns, which is why option B accurately reflects their concern. Option D is incorrect because the PR description itself focuses on API updates, not fairness evaluation.
8 / 25
Senior Dev: 'Okay team, I've been reviewing the output from the new recommendation engine. The API response is returning these scores, but frankly, I don't understand *why* it's recommending this particular product to this user – there's no clear traceability back to their profile or past interactions. It feels like a black box! PR Description: 'This PR implements the new recommendation engine, utilizing collaborative filtering for personalized suggestions.' Which of the following best describes the issue highlighted by the Senior Dev?The Senior Dev is concerned about the lack of transparency within the recommendation engine – a common problem when models are complex. Opaque refers to something that's difficult or impossible to understand, and in this context, it describes the model's decision-making process. The PR description focuses on implementation, not the crucial aspect of explainability, highlighting the need for a system where users can see *why* a recommendation was made.
9 / 25
Reviewer: 'We've noticed a significant correlation between the model's output and certain protected characteristics – specifically, higher rates of denial for loan applications submitted by applicants from minority ethnic groups. While we've documented steps taken to address data bias during training, the model's decision-making process remains largely opaque. PR Description: 'This PR implements a new fairness metric and integrates it into the model evaluation pipeline.' Which phrase most accurately reflects the reviewer's concern?The reviewer is highlighting a critical issue: algorithmic bias. The core problem isn't simply about accuracy (option A), but about whether the model is producing *fair* outcomes across different demographic groups. 'Opaque' (option B) perfectly captures the concern that the decision-making process is not transparent, making it difficult to identify and correct potential biases—this is a key aspect of AI transparency. The PR focuses on metrics; the reviewer needs more than just measurement – they need insight into *why* the model is behaving in this way (option C). Simply updating metrics doesn't address the underlying lack of explainability (option D).
10 / 25
During a Slack discussion about the new customer churn prediction model, Sarah writes: 'Just ran the explainability dashboard – it's showing that users flagged for potential churn are overwhelmingly associated with low engagement scores and infrequent logins. However, the *reason* why those low engagement scores are being predicted is… unclear. It just says 'model confidence.' This feels like we're building a system without understanding *why* it's making these decisions.' Which of the following best describes Sarah's concern?
Sarah's concern isn't simply that the model lacks justification; she highlights the issue of 'model confidence' being presented as an explanation. The incorrect options assume a direct explanation or that confidence alone is enough – Sarah wants to know *why* engagement scores are low and *how* that leads to churn prediction, not just that the model thinks it's likely. A good explanation dashboard would provide root cause analysis beyond simple confidence levels.
11 / 25
Reviewer: 'I'm seeing a high degree of disagreement between the model's predictions and the actual outcomes in this fraud detection cohort. The explanations are also quite opaque – it's hard to understand *why* the model flagged these transactions as suspicious. PR Description: 'This PR implements the new fraud detection model, leveraging a gradient boosting algorithm for improved accuracy.'Which of the following best describes the reviewer's concern regarding this PR?
The reviewer is highlighting a lack of transparency, specifically referring to 'opaque' decision-making. This term describes models where the reasoning behind predictions isn't clear or easily understood – often due to complex algorithms like gradient boosting. The options presented focus on different potential issues (overfitting, bias, insufficient documentation), but only 'opaque' directly addresses the core concern of difficult-to-interpret model behavior. It's crucial for developers to be able to articulate this lack of transparency when discussing models with stakeholders and reviewers.
12 / 25
Reviewer: 'The model's output is consistently exhibiting bias across different demographic groups. While the technical documentation describes the training data preprocessing steps, it lacks a clear articulation of how potential biases were identified and mitigated – essentially, there's no transparency into the fairness evaluation process. PR Description: 'This PR updates the API endpoint to include confidence scores for each prediction.' Which phrase best captures the reviewer's primary concern?The reviewer isn't primarily focused on the technical details of the API documentation (option A) or the accuracy of the confidence scores (option C). Instead, they are highlighting a critical lack of transparency regarding bias mitigation. The core issue is the absence of documented strategies for identifying and addressing potential biases within the model's training data and outputs – this directly relates to fairness concerns, which is why option B accurately reflects their concern. Option D is incorrect because the PR description itself focuses on API updates, not fairness evaluation.
13 / 25
Senior Dev: 'Okay team, I've been reviewing the output from the new recommendation engine. The API response is returning these scores, but frankly, I don't understand *why* it's recommending this particular product to this user – there's no clear traceability back to their profile or past interactions. It feels like a black box! PR Description: 'This PR implements the new recommendation engine, utilizing collaborative filtering for personalized suggestions.' Which of the following best describes the issue highlighted by the Senior Dev?The Senior Dev is concerned about the lack of transparency within the recommendation engine – a common problem when models are complex. Opaque refers to something that's difficult or impossible to understand, and in this context, it describes the model's decision-making process. The PR description focuses on implementation, not the crucial aspect of explainability, highlighting the need for a system where users can see *why* a recommendation was made.
14 / 25
Reviewer: 'We've noticed a significant correlation between the model's output and certain protected characteristics – specifically, higher rates of denial for loan applications submitted by applicants from minority ethnic groups. While we've documented steps taken to address data bias during training, the model's decision-making process remains largely opaque. PR Description: 'This PR implements a new fairness metric and integrates it into the model evaluation pipeline.' Which phrase most accurately reflects the reviewer's concern?The reviewer is highlighting a critical issue: algorithmic bias. The core problem isn't simply about accuracy (option A), but about whether the model is producing *fair* outcomes across different demographic groups. 'Opaque' (option B) perfectly captures the concern that the decision-making process is not transparent, making it difficult to identify and correct potential biases—this is a key aspect of AI transparency. The PR focuses on metrics; the reviewer needs more than just measurement – they need insight into *why* the model is behaving in this way (option C). Simply updating metrics doesn't address the underlying lack of explainability (option D).
15 / 25
During a Slack discussion about the new customer churn prediction model, Sarah writes: 'Just ran the explainability dashboard – it's showing that users flagged for potential churn are overwhelmingly associated with low engagement scores and infrequent logins. However, the *reason* why those low engagement scores are being predicted is… unclear. It just says 'model confidence.' This feels like we're building a system without understanding *why* it's making these decisions.' Which of the following best describes Sarah's concern?
Sarah's concern isn't simply that the model lacks justification; she highlights the issue of 'model confidence' being presented as an explanation. The incorrect options assume a direct explanation or that confidence alone is enough – Sarah wants to know *why* engagement scores are low and *how* that leads to churn prediction, not just that the model thinks it's likely. A good explanation dashboard would provide root cause analysis beyond simple confidence levels.
16 / 25
Reviewer: 'I'm seeing a high degree of disagreement between the model's predictions and the actual outcomes in this fraud detection cohort. The explanations are also quite opaque – it's hard to understand *why* the model flagged these transactions as suspicious. PR Description: 'This PR implements the new fraud detection model, leveraging a gradient boosting algorithm for improved accuracy.'Which of the following best describes the reviewer's concern regarding this PR?
The reviewer is highlighting a lack of transparency, specifically referring to 'opaque' decision-making. This term describes models where the reasoning behind predictions isn't clear or easily understood – often due to complex algorithms like gradient boosting. The options presented focus on different potential issues (overfitting, bias, insufficient documentation), but only 'opaque' directly addresses the core concern of difficult-to-interpret model behavior. It's crucial for developers to be able to articulate this lack of transparency when discussing models with stakeholders and reviewers.
17 / 25
Reviewer: 'The model's output is consistently exhibiting bias across different demographic groups. While the technical documentation describes the training data preprocessing steps, it lacks a clear articulation of how potential biases were identified and mitigated – essentially, there's no transparency into the fairness evaluation process. PR Description: 'This PR updates the API endpoint to include confidence scores for each prediction.' Which phrase best captures the reviewer's primary concern?The reviewer isn't primarily focused on the technical details of the API documentation (option A) or the accuracy of the confidence scores (option C). Instead, they are highlighting a critical lack of transparency regarding bias mitigation. The core issue is the absence of documented strategies for identifying and addressing potential biases within the model's training data and outputs – this directly relates to fairness concerns, which is why option B accurately reflects their concern. Option D is incorrect because the PR description itself focuses on API updates, not fairness evaluation.
18 / 25
Senior Dev: 'Okay team, I've been reviewing the output from the new recommendation engine. The API response is returning these scores, but frankly, I don't understand *why* it's recommending this particular product to this user – there's no clear traceability back to their profile or past interactions. It feels like a black box! PR Description: 'This PR implements the new recommendation engine, utilizing collaborative filtering for personalized suggestions.' Which of the following best describes the issue highlighted by the Senior Dev?The Senior Dev is concerned about the lack of transparency within the recommendation engine – a common problem when models are complex. Opaque refers to something that's difficult or impossible to understand, and in this context, it describes the model's decision-making process. The PR description focuses on implementation, not the crucial aspect of explainability, highlighting the need for a system where users can see *why* a recommendation was made.
19 / 25
Reviewer: 'We've noticed a significant correlation between the model's output and certain protected characteristics – specifically, higher rates of denial for loan applications submitted by applicants from minority ethnic groups. While we've documented steps taken to address data bias during training, the model's decision-making process remains largely opaque. PR Description: 'This PR implements a new fairness metric and integrates it into the model evaluation pipeline.' Which phrase most accurately reflects the reviewer's concern?The reviewer is highlighting a critical issue: algorithmic bias. The core problem isn't simply about accuracy (option A), but about whether the model is producing *fair* outcomes across different demographic groups. 'Opaque' (option B) perfectly captures the concern that the decision-making process is not transparent, making it difficult to identify and correct potential biases—this is a key aspect of AI transparency. The PR focuses on metrics; the reviewer needs more than just measurement – they need insight into *why* the model is behaving in this way (option C). Simply updating metrics doesn't address the underlying lack of explainability (option D).
20 / 25
During a Slack discussion about the new customer churn prediction model, Sarah writes: 'Just ran the explainability dashboard – it's showing that users flagged for potential churn are overwhelmingly associated with low engagement scores and infrequent logins. However, the *reason* why those low engagement scores are being predicted is… unclear. It just says 'model confidence.' This feels like we're building a system without understanding *why* it's making these decisions.' Which of the following best describes Sarah's concern?
Sarah's concern isn't simply that the model lacks justification; she highlights the issue of 'model confidence' being presented as an explanation. The incorrect options assume a direct explanation or that confidence alone is enough – Sarah wants to know *why* engagement scores are low and *how* that leads to churn prediction, not just that the model thinks it's likely. A good explanation dashboard would provide root cause analysis beyond simple confidence levels.
21 / 25
Reviewer: 'I'm seeing a high degree of disagreement between the model's predictions and the actual outcomes in this fraud detection cohort. The explanations are also quite opaque – it's hard to understand *why* the model flagged these transactions as suspicious. PR Description: 'This PR implements the new fraud detection model, leveraging a gradient boosting algorithm for improved accuracy.'Which of the following best describes the reviewer's concern regarding this PR?
The reviewer is highlighting a lack of transparency, specifically referring to 'opaque' decision-making. This term describes models where the reasoning behind predictions isn't clear or easily understood – often due to complex algorithms like gradient boosting. The options presented focus on different potential issues (overfitting, bias, insufficient documentation), but only 'opaque' directly addresses the core concern of difficult-to-interpret model behavior. It's crucial for developers to be able to articulate this lack of transparency when discussing models with stakeholders and reviewers.
22 / 25
Reviewer: 'The model's output is consistently exhibiting bias across different demographic groups. While the technical documentation describes the training data preprocessing steps, it lacks a clear articulation of how potential biases were identified and mitigated – essentially, there's no transparency into the fairness evaluation process. PR Description: 'This PR updates the API endpoint to include confidence scores for each prediction.' Which phrase best captures the reviewer's primary concern?The reviewer isn't primarily focused on the technical details of the API documentation (option A) or the accuracy of the confidence scores (option C). Instead, they are highlighting a critical lack of transparency regarding bias mitigation. The core issue is the absence of documented strategies for identifying and addressing potential biases within the model's training data and outputs – this directly relates to fairness concerns, which is why option B accurately reflects their concern. Option D is incorrect because the PR description itself focuses on API updates, not fairness evaluation.
23 / 25
Senior Dev: 'Okay team, I've been reviewing the output from the new recommendation engine. The API response is returning these scores, but frankly, I don't understand *why* it's recommending this particular product to this user – there's no clear traceability back to their profile or past interactions. It feels like a black box! PR Description: 'This PR implements the new recommendation engine, utilizing collaborative filtering for personalized suggestions.' Which of the following best describes the issue highlighted by the Senior Dev?The Senior Dev is concerned about the lack of transparency within the recommendation engine – a common problem when models are complex. Opaque refers to something that's difficult or impossible to understand, and in this context, it describes the model's decision-making process. The PR description focuses on implementation, not the crucial aspect of explainability, highlighting the need for a system where users can see *why* a recommendation was made.
24 / 25
Reviewer: 'We've noticed a significant correlation between the model's output and certain protected characteristics – specifically, higher rates of denial for loan applications submitted by applicants from minority ethnic groups. While we've documented steps taken to address data bias during training, the model's decision-making process remains largely opaque. PR Description: 'This PR implements a new fairness metric and integrates it into the model evaluation pipeline.' Which phrase most accurately reflects the reviewer's concern?The reviewer is highlighting a critical issue: algorithmic bias. The core problem isn't simply about accuracy (option A), but about whether the model is producing *fair* outcomes across different demographic groups. 'Opaque' (option B) perfectly captures the concern that the decision-making process is not transparent, making it difficult to identify and correct potential biases—this is a key aspect of AI transparency. The PR focuses on metrics; the reviewer needs more than just measurement – they need insight into *why* the model is behaving in this way (option C). Simply updating metrics doesn't address the underlying lack of explainability (option D).
25 / 25
During a Slack discussion about the new customer churn prediction model, Sarah writes: 'Just ran the explainability dashboard – it's showing that users flagged for potential churn are overwhelmingly associated with low engagement scores and infrequent logins. However, the *reason* why those low engagement scores are being predicted is… unclear. It just says 'model confidence.' This feels like we're building a system without understanding *why* it's making these decisions.' Which of the following best describes Sarah's concern?
Sarah's concern isn't simply that the model lacks justification; she highlights the issue of 'model confidence' being presented as an explanation. The incorrect options assume a direct explanation or that confidence alone is enough – Sarah wants to know *why* engagement scores are low and *how* that leads to churn prediction, not just that the model thinks it's likely. A good explanation dashboard would provide root cause analysis beyond simple confidence levels.
Frequently Asked Questions
What will I practice in "AI Transparency Vocabulary"?
This is an AI Ethics exercise set. It walks through 25 scenario-based multiple-choice questions built around real usage of AI Ethics terminology that IT professionals encounter on the job.
Is this exercise free to use?
Yes. Every exercise on CoderSlingo, including this one, is free to complete with no account, sign-up, or paywall.
How many questions are in this exercise?
This set contains 25 questions. Each one shows immediate feedback and a detailed explanation after you answer, so you learn the correct usage right away rather than waiting for a final score.
Do I need prior experience to complete this exercise?
No prior experience is required. Each question includes a full explanation covering the reasoning behind the correct answer, so the exercise itself teaches the AI Ethics vocabulary as you go.
Can I retry the exercise if I get questions wrong?
Yes — use the "Try again" button on the results screen to reset your answers and go through all the questions again. There is no limit on attempts.
Is my progress saved?
Your answers and score for the current session are tracked in the browser as you go. No account or login is needed, and there is nothing to install.
What if I don't understand a term used in a question?
Read the explanation shown after you answer each question — it breaks down the correct term in plain English with a real-world example. You can also check the site Glossary for quick definitions.
How is this different from reading a blog article on the topic?
Exercises like this one are interactive drills that test and reinforce specific vocabulary through multiple-choice questions, while blog articles explain concepts in prose. Practising here after reading builds active recall, not just passive recognition.
Where can I find more AI Ethics exercises?
See the AI Ethics exercises hub for the full set of related pages, or browse all exercise categories from the main Exercises index.
Can I use this exercise to prepare for a technical interview?
Yes — AI Ethics vocabulary comes up often in technical discussions and interviews. Pair this exercise with our dedicated Interview Preparation section for role-specific practice.