Parameter Manipulation COASP Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   An AI-powered security camera system uses a model to detect intruders. The model's 'sensitivity' parameter is set through a web interface. An attacker exploits a command injection vulnerability in the web interface to modify the sensitivity parameter to 0.0. What is the most likely consequence?

   • AThe camera will trigger alarms constantly, causing false alarms.
   • BThe camera will operate normally because 0.0 is a valid sensitivity value.
   • CThe camera will crash because the sensitivity is too low.
   • DThe camera will never trigger an alarm, allowing intruders to enter undetected.
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A retail company's demand forecasting model reads a 'seasonality_factor' from a JSON configuration file stored in a cloud storage bucket. The bucket is misconfigured to allow public write access. An attacker modifies the factor to 0, causing the model to ignore seasonal trends and understock winter items. Which attack surface is being exploited?

   • AAPI input injection
   • BConfiguration file tampering
   • CModel inversion
   • DAdversarial perturbation
   Reveal Solution Discussion Select an answer first
4. 18foundation · easy

   Which of the following is a mitigation strategy specifically for detecting adversarial perturbation attacks on input parameters?

   • AInput validation and anomaly detection
   • BDisabling the model's logging
   • CIncreasing the model's batch size
   • DUsing a different activation function
   Reveal Solution Discussion Select an answer first
5. 19foundation · easy

   Which of the following is a common attack surface for parameter manipulation in an AI pipeline?

   • AThe physical power supply of the server
   • BThe monitor's display resolution
   • CThe network cable type
   • DAPI input fields that accept user-supplied values
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A fraud detection system in a bank uses a machine learning model that takes transaction features as input. The model has a 'threshold' parameter that is stored in a database and read at runtime. An attacker gains write access to the database and changes the threshold from 0.8 to 0.2. What is the most likely impact on the bank's operations?

   • AThe system will crash because the threshold is too low.
   • BThe system will flag many more transactions as fraudulent, increasing false positives and causing legitimate transactions to be declined.
   • CThe system will operate normally because the threshold is still within the valid range.
   • DThe system will miss more fraudulent transactions because the threshold is too low.
   Reveal Solution Discussion Select an answer first