Which of the following is a PRIMARY consideration to protect against privacy violations when utilizing artificial intelligence (AI) driven business decisions?

Question

Options

• ADe-identifying the data to be analyzed
• BVerifying the data subjects have consented to the processing
• CDefining the intended objectives
• DEnsuring proper data sets are used to train the models

• A
  2% (1)
• B
  7% (4)
• C
  11% (6)
• D
  80% (43)

Explanation

Ensuring proper data sets are used to train AI models is the primary privacy consideration because the training data is foundational to every downstream decision the model makes. If a model is trained on improperly sourced, biased, or PII-laden data, it will systematically reproduce privacy violations at scale - a problem unique to AI systems. Poor training data can cause models to memorize and inadvertently expose personal information, or make discriminatory inferences about individuals.

Why the distractors fall short:

• A (De-identifying data) is a useful technique but secondary - de-identification can be reversed, and it doesn't address whether the underlying training data was appropriate in the first place.
• B (Verifying consent) matters for legal compliance, but original consent may not cover the new AI use case, and consent alone doesn't prevent a poorly trained model from making privacy-invasive decisions.
• C (Defining intended objectives) is a governance step, not a privacy control - you can have clear objectives and still train a model on problematic data.

Memory tip: Think GIGO - Garbage In, Garbage Out. In AI, "garbage" includes improperly sourced or privacy-exposing training data. The dataset is the root cause; every other control is downstream of it.

No community discussion yet for this question.