An organization is collecting data to train a generative AI model for customer service. They want to ensure security throughout the ML lifecycle. What is a critical consideration at this stage?

Question

Options

• AImplementing access controls and protecting sensitive information within the training data.
• BApplying the latest software patches to the AI model on a regular basis.
• CEstablishing ethical guidelines for AI model responses to ensure fairness and avoid harm.
• DMonitoring the AI model's performance for unexpected outputs and potential errors.

• A
  92% (36)
• C
  5% (2)
• D
  3% (1)

Explanation

At the data collection stage of the ML lifecycle, the critical security consideration is implementing access controls and protecting sensitive information within the training data (A). Training data for customer service often contains personally identifiable information (PII), account details, or confidential interactions - a breach at this stage contaminates the entire pipeline. Applying software patches (B) is an operational security task unrelated to training data collection. Establishing ethical AI guidelines (C) is important but is a governance/design concern, not a data-collection security control. Monitoring model outputs for errors (D) occurs post-deployment, not during data collection. Data-level security is the most critical concern at this specific lifecycle stage.

No community discussion yet for this question.