Skip to content
AI Security

Interview question · Adversarial ML & Model Privacy

What is transferability of adversarial examples, and why doesn't hiding a model behind an API stop adversarial attacks?

An attacker who cannot see your model can attack one they can see, and much of the damage carries over. That single property turns 'the weights are secret' into a weak defence.

mediumNewPreparation tags: Google DeepMind, Microsoft, MetaUpdated Nov 2026

An attacker who cannot see your model can attack one they can see, and much of the damage carries over. That single property turns 'the weights are secret' into a weak defence.

Make your first attempt

What would you say before reading the answer?

  1. 01 State the assumptions your answer depends on.
  2. 02 Explain the mechanism and choose a control.
  3. 03 Name a condition that would change your decision.

The worked answer includes an approach, scoring criteria and follow-up questions. Compare the reasoning with your own.

10 answers per topic instead of 5, plus saved progress and bookmarks · costs nothingor unlock all 529 remaining answers · $25

Every worked answer in all 17 topics, every lesson in every course, the full concept curriculum, the complete practice-test bank, and the downloadable security field guides.

6 months · One payment · No auto-renewal

What if · the follow-ups in this answer

The answer works through 6 follow-ups like an interviewer would

A team assumes attackers cannot craft adversarial examples against its document classifier because the model is private and only labels are returned.

  1. What is transferability?
  2. How would an attacker use it?
  3. Does this apply to language models?
  4. What reduces transfer success?
  5. How would you test for transfer attacks?
  6. What would you tell the team?

Each rung has a nudge, a model answer, the principle that carries over and the weak answer to avoid. Premium opens every answer and ladder.

UP NEXT ON YOUR JOURNEY
Next in this trackHow do adversarial images attack vision-language models, how is that different from text in images, and does preprocessing help?Next in this trackHow would you test whether a model update increased memorisation, and does a leak that needs thousands of queries matter?Next in this trackHow would you red-team a content moderation classifier for evasion, and report results that are useful to the team that owns it?
DISCUSSION · 0

No comments yet. Share a correction or a different way to answer it.