Artificial Intelligence Governance Professional AIGP · Free Practice Question Medium
Question 8
A transportation app company realizes its ride-fare prediction AI is consistently underestimating fares in specific neighborhoods. Investigation reveals that large numbers of fake, low-fare rides were injected into the training dataset from a particular source. What type of threat does this scenario exemplify?
- A Natural data drift due to changing city dynamics
- B Overfitting to rare cases in the data
- C A data leak exposing fare data to the public
- D Data poisoning attack, deliberately introducing incorrect data into training
Reveal correct answer
Correct answer: D
Explanation
Deliberate injection of misleading data for model corruption is the hallmark of a data poisoning attack.A. Data drift happens naturally, not through deliberate tampering.
B. Overfitting is a modeling issue, not a result of adversarial input.
C. A data leak is accidental, not adversarial data manipulation.
D. This is a classic example of data poisoning, where malicious input manipulates the model's outputs.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
You must be logged in to post a comment.
