IAPP Updated AIGP Exam Questions and Answers by iona

The common types of machine learning include supervised learning, unsupervised learning, reinforcement learning, and deep learning. Cognitive learning is not a type of machine learning; rather, it is a term often associated with the broader field of cognitive science and psychology. Reference: AIGP BODY OF KNOWLEDGE and standard AI/ML literature.

The correct answer isD. GPS location data isnot biometric data—it is consideredgeolocation data, which is personal data butnot biometricunder most U.S. laws.

From the AIGP ILT Guide (Data Privacy Module):

“Biometric data includes measurable biological or behavioral characteristics such as iris scans, facial recognition, voice prints, and keystroke patterns when used for identification.”

AI Governance in Practice Report 2025 (Privacy and Data Protection section):

“Location data, while sensitive, is not considered biometric unless it’s tied to a uniquely identifyingbiological trait.”

Thus,GPS location data, while potentially sensitive, is not classified as biometric.

To evaluate the effectiveness of an AI-based ad recommendation tool, the most relevant metric is the output data, specifically assessing the delta between the prediction and actual ad clicks. This metric directly measures the tool ' s accuracy and effectiveness in making accurate recommendations that lead to user engagement. While monitoring algorithmic patterns and input data can provide insights into the model ' s behavior and targeting accuracy, and GPU performance can indicate the robustness and efficiency of the tool, the primary indicator of effectiveness for an ad recommendation tool is how well it predicts actual ad clicks.

AI governance frameworks consistently define the “unique characteristics” of AI that create novel governance challenges.

Across the OECD AI Principles, NIST AI RMF, ISO/IEC 42001, and the EU AI Act, the key characteristics requiring governance are:

Autonomy – AI systems act without direct human intervention and take context-dependent decisions.

Adaptability – AI systems learn, evolve, and update their behavior over time, making outputs non-deterministic.

These traits fundamentally distinguish AI from traditional software and shape all major governance activities (risk assessment, monitoring, assurance, accountability, alignment, etc.).

Below is why each option is correct or incorrect:

A. Autonomy — NOT selected

Reason:

Autonomy is repeatedly cited as one of the defining characteristics that necessitates governance.

NIST AI RMF references the risks of autonomous behavior .

ISO/IEC 42001 emphasizes governance controls for “systems operating with varying levels of autonomy.”

Thus, Autonomy IS a unique characteristic , so it is not an answer.

B. Automation — SELECTED

Reason:

“Automation” is not a unique AI property. Automation predates AI and applies to robotics, scripts, business process automation, etc.

AI governance literature does not classify automation as a unique AI-specific characteristic.

Automation is a result of AI, but not a distinguishing property .

Therefore, Automation is correctly selected as NOT a unique AI characteristic.

C. Adaptability — NOT selected

Reason:

Adaptability (systems that update, learn, or change behavior) is recognized universally as a core distinctive trait of AI.

EU AI Act emphasizes governance requirements for “adaptive systems.”

NIST AI RMF highlights “learning and emergent behavior.”

Thus, Adaptability IS a unique characteristic , so it should not be chosen as an exception.

D. Speed and scale — SELECTED

Reason:

While AI can operate at high speed and scale, these are not unique characteristics of AI .

Traditional computing systems, cloud workloads, and automation pipelines operate at similar scale.

AI governance documents mention speed/scale as a risk amplifier , not a defining characteristic.

Therefore, Speed and scale should be selected.

E. Superintelligence — SELECTED

Reason:

“Superintelligence” is not a characteristic of current AI systems and is not referenced as a governance-relevant attribute in mainstream policy documents (OECD, NIST, ISO, EU AI Act).

It is speculative rather than a defining feature.

Therefore, it is correctly chosen as NOT a unique characteristic requiring governance today.