Computer — All MCQs
4795 questions available
Q2081.Which AI risk specifically targets the training pipeline by injecting malicious data?
Correct Answer: A
Data poisoning corrupts training data to manipulate model behavior.
Q2082.Which technique would help detect AI-generated text in fraud cases?
Correct Answer: A
Specialized classifiers attempt to distinguish AI-generated from human text.
Q2083.Which is TRUE about 'embeddings'?
Correct Answer: A
Embeddings represent data as dense vectors that capture semantic relationships.
Q2084.Which AI capability supports real-time license-plate recognition at checkpoints?
Correct Answer: A
Automatic Number Plate Recognition (ANPR) uses CV and OCR to read plates in real time.
Q2085.Which statement about 'AGI' is correct as of current technology?
Correct Answer: D
Current systems are narrow/advanced AI; true human-level AGI remains unachieved.
Q2086.Which approach reduces hallucinations by letting the model cite verifiable sources?
Correct Answer: B
Grounding responses in cited sources reduces unsupported hallucinations.
Q2087.Which is the BEST AI use case for FIA cybercrime wing analyzing seized phones?
Correct Answer: D
AI can extract data, link entities, and rebuild event timelines from seized devices.
Q2088.Which metric combines precision and recall into a single value?
Correct Answer: D
The F1-score is the harmonic mean of precision and recall.
Q2089.Which AI technique can de-anonymize patterns in metadata for investigations?
Correct Answer: C
Graph/network analysis reveals relationships and patterns in communication metadata.
Q2090.Which is a key challenge of deploying AI evidence in court?
Correct Answer: C
Courts require explainable, reliable, and admissible AI-derived evidence.
Q2091.An FIA analyst must trace laundered funds across thousands of crypto wallets. The MOST suitable AI approach is:
Correct Answer: C
Graph analysis with address clustering traces fund flows across crypto wallets.
Q2092.A suspect submits a video alibi. AI analysis reveals inconsistent blinking and lip-sync artifacts. This indicates:
Correct Answer: D
Inconsistent blinking and lip-sync artifacts are common deepfake indicators.
Q2093.Which combination best supports FIA in automated CNIC/passport verification at borders?
Correct Answer: C
OCR reads documents, facial recognition verifies identity, and DB matching confirms records.
Q2094.An AI fraud model flags many innocent users (high false positives). To fix this WITHOUT missing real fraud, an analyst should primarily:
Correct Answer: B
Threshold tuning and better features can raise precision while preserving recall.
Q2095.Which is the GREATEST ethical risk of predictive policing systems?
Correct Answer: B
Predictive policing can amplify historical biases embedded in past crime data.
Q2096.An investigator needs to verify whether a confession letter was AI-generated. The BEST first step is:
Correct Answer: A
AI-text detectors and stylometry help assess whether text was machine-generated.
Q2097.Which scenario BEST illustrates a 'data poisoning' attack on an FIA fraud-detection model?
Correct Answer: A
Injecting mislabeled data during training corrupts the model to evade detection later.
Q2098.To analyze millions of intercepted messages for threats while preserving privacy across agencies, the BEST approach is:
Correct Answer: C
Federated learning enables collaborative analysis without centralizing raw private data.
Q2099.An LLM assistant gives an investigator a fabricated legal citation. The correct professional response is to:
Correct Answer: D
AI outputs, especially citations, must be independently verified due to hallucination risk.
Q2100.Which AI technique would best reconstruct a suspect's movement timeline from CDRs and GPS logs?
Correct Answer: D
Spatiotemporal/sequence analysis reconstructs movement timelines from location data.