GENERATIVE-AI-LEADER · Question #78
A global travel booking platform named VistaVoyage is developing a generative AI system to identify payment fraud across about 45 million reservations each day. The team is concerned that…
The correct answer is C. It should be continuous with robustness techniques embedded during model training and. Adversarial robustness must be embedded continuously across the entire ML lifecycle-including threat modeling, data collection, preprocessing, model training (adversarial training, regularization), evaluation, deployment, and ongoing monitoring. Limiting protections to only the…
Question
A global travel booking platform named VistaVoyage is developing a generative AI system to identify payment fraud across about 45 million reservations each day. The team is concerned that adversaries may make small tweaks to inputs so the model incorrectly treats fraudulent behavior as legitimate. At what point in the machine learning lifecycle should robust protections against these adversarial tactics be established to preserve security?
Options
- ALimited to the business requirements and initial threat modeling stage
- BHandled mostly with input sanitation and validation in Dataflow pipelines before training or serving
- CIt should be continuous with robustness techniques embedded during model training and
- DExclusively when the model is released to production
How the community answered
(43 responses)- A16% (7)
- B5% (2)
- C72% (31)
- D7% (3)
Explanation
Adversarial robustness must be embedded continuously across the entire ML lifecycle-including threat modeling, data collection, preprocessing, model training (adversarial training, regularization), evaluation, deployment, and ongoing monitoring. Limiting protections to only the business requirements phase (A), only input validation pipelines (B), or only production release (D) leaves large windows of vulnerability. Adversaries can exploit weaknesses introduced at any lifecycle stage, so defenses must be present at every stage.
Topics
Community Discussion
No community discussion yet for this question.