IBM A1000-068 Practice Exam 2025: Latest Questions

A team is explaining why a model performs well on the training data but poorly on new customer records. Which concept best describes this situation?

An insurance analyst wants an AI system to automatically label incoming emails into categories such as 'claim status', 'policy change', and 'billing question'. Which type of machine learning problem is this?

A bank wants to understand which input factors most influenced an AI decision to approve or deny a loan, and to be able to explain this to auditors. Which capability is MOST relevant?

A developer needs to add natural-language understanding and text generation to an internal application while keeping governance controls and deployment options suitable for enterprise use. Which IBM offering is most aligned with this need?

A retail company trained a demand-forecasting model and achieved strong results on a holdout set. After deployment, accuracy degrades over several weeks as buying patterns change. What is the MOST likely cause?

A team wants to build a generative AI assistant for employees. The assistant must answer using internal policy documents and reduce hallucinations by grounding responses in trusted sources. Which approach is MOST appropriate?

A healthcare organization is creating an AI model using sensitive patient data. Which practice BEST supports responsible AI and compliance while enabling model development?

A team is deciding between batch scoring and real-time scoring for an AI model. Fraud detection needs an immediate decision during a transaction, while daily marketing segmentation can run overnight. What is the BEST deployment choice?

A company wants to deploy a generative AI feature that summarizes customer support tickets. They are concerned the model may output personal data from training or retrieved documents. Which combination of controls is MOST effective to reduce this risk?

An enterprise is building an AI solution that must be reusable across multiple departments. Requirements include: consistent model deployment, experiment tracking, audit-friendly governance, and the ability to update models with minimal downtime. Which architecture pattern BEST meets these needs?

A team is trying to explain the difference between supervised and unsupervised learning to business stakeholders. Which statement is MOST accurate?

A developer wants to add a natural-language chatbot to a customer support portal with minimal ML expertise. Which IBM technology is MOST appropriate to build the conversational assistant?

A project team is documenting model performance for a classification model. Which metric is MOST appropriate when the positive class is rare and false negatives are especially costly?

A retail company wants to answer questions using internal policy documents. They need the response to cite relevant passages and reduce hallucinations. Which approach is MOST appropriate?

A data scientist reports a model shows 98% accuracy in production, but customer complaints indicate it misses many fraud cases. The dataset contains 2% fraud. What is the MOST likely issue?

A company wants to operationalize an ML model with consistent deployments, model versioning, and approval gates for promotion to production. Which capability BEST supports this need in an IBM AI platform context?

A bank is required to explain adverse credit decisions to customers. Which practice is MOST appropriate to support explainability and regulatory expectations?

A team deployed a text classification model and notices performance has degraded over time. The input text now contains new slang and abbreviations. What is the MOST likely cause and best next step?

A healthcare organization wants to use an LLM to summarize clinician notes. They must minimize the risk of exposing sensitive personal information during development and testing. Which approach is MOST appropriate?

A company is building a generative AI assistant for internal employees. They need to reduce harmful or policy-violating outputs and demonstrate controls to auditors. Which combination is MOST effective?

A business analyst wants to extract key fields (invoice number, total amount, vendor name) from thousands of scanned invoice PDFs with minimal custom model training. Which IBM capability best fits this requirement?

A team built a high-accuracy model, but customer complaints indicate it performs poorly for a small, underrepresented user segment. Which action is the BEST first step to address the issue?

A retailer wants to reduce customer churn. The dataset includes customer demographics, purchase history, support interactions, and a label indicating whether the customer churned. Which type of machine learning problem is this?

A healthcare organization wants to use an LLM to summarize clinician notes. They must ensure sensitive data is protected and that model outputs can be audited for risky content. Which combination is MOST appropriate?

A team deployed a model and notices accuracy gradually degrading over several weeks, even though the application code has not changed. Which issue is the MOST likely cause and what is the best response?