IBM A1000-075: Foundations of AI Practice Exam 2025: Latest Questions

A retail team wants a quick way to describe what AI does to non-technical stakeholders. Which statement best describes machine learning?

A developer is building a chatbot that must answer questions using information from internal policy documents. The team wants a managed IBM service designed for conversational experiences. Which IBM Watson capability best fits?

A data scientist is evaluating a binary classification model for detecting fraudulent transactions. Fraud is rare, and missing fraud is expensive. Which metric is often more informative than accuracy in this scenario?

An organization wants to ensure an AI system’s decisions can be explained to affected users and auditors. Which ethical principle does this most directly support?

A team built a sentiment classifier and sees strong performance in testing, but performance drops sharply after deployment because customer language changes over time. What is the most likely issue?

A bank wants to reduce the risk of unintentionally training on sensitive customer attributes (e.g., health status) and to track who can access curated datasets used for AI. Which IBM capability best supports this governance need?

A product team wants to experiment with multiple model types and track experiments (data, parameters, and results) in a repeatable way. Which approach is a best practice?

A company is deploying an AI-powered customer support assistant. They want to reduce harm from incorrect or inappropriate outputs while still improving over time. Which governance control is most appropriate?

A team is building a model to approve loans. They discover the training data contains historically lower approval rates for a protected group. They want to reduce discriminatory outcomes while keeping the model useful. Which action is most appropriate as an initial mitigation step?

An AI solution must be designed so that a conversational interface can use enterprise documents while preventing disclosure of confidential content to unauthorized users. Which architecture pattern best addresses this requirement?

A project team is discussing AI terms. Which statement best describes the relationship between AI, machine learning (ML), and deep learning (DL)?

A bank wants an easy way to turn common customer questions into a chat experience on its website, with intent recognition and dialog flow. Which IBM capability best fits this need?

You are preparing a dataset for a supervised learning model. Which choice correctly pairs a problem type with its typical target variable?

A team is evaluating model performance on a dataset where only 2% of transactions are fraudulent. Which metric is generally more informative than accuracy for this scenario?

A product owner asks why the model performed well during training but poorly on new data. Which issue most directly explains this behavior?

An HR team wants to use AI to screen resumes. Which governance practice best helps reduce the risk of unfair bias in the model’s decisions?

A developer is building an app that generates summaries of internal policy documents. The team wants to reduce the chance of the model fabricating content not supported by the documents. Which approach is most appropriate?

A team deployed a model and notices performance degradation over time as customer behavior changes. What is the BEST next step?

A healthcare organization wants to use a foundation model for summarizing clinical notes. Which control is MOST critical to meet privacy and compliance expectations?

A team uses an automated pipeline to train a model. They scale numeric features using statistics computed on the full dataset (train + test) before splitting. Test results look unusually high. What is the MOST likely cause?

A retail team wants to quickly understand the main themes in 5,000 open-ended customer survey comments and identify the most common issues without building a custom ML model. Which approach best fits this requirement?

A bank is building an AI system to assist call-center agents. The model sometimes provides confident-sounding answers that are not supported by policy documents. Which governance control most directly reduces this risk?

A data scientist evaluates a binary classifier for fraud detection. Fraud cases are rare, and missing fraud is costly. Which metric is generally most appropriate to focus on first in this situation?

A team is building a model to predict equipment failure. They include a feature called "days_since_last_service". During training, they accidentally compute this value using the actual service date that occurred after the failure for some records. What is the primary issue with this dataset?

Which statement best describes the difference between artificial intelligence (AI) and machine learning (ML)?