Ethical, Societal, and Security Concerns in AI AIE Practice Questions (Page 6)

Questions 26–30

1. 26application · medium

   A financial institution uses an AI model to detect fraudulent transactions. The model is updated daily with new transaction data. Recently, the model's false positive rate has increased, causing legitimate customers to be declined. The institution suspects that attackers are injecting fraudulent transactions into the training data to make the model overly cautious. The institution wants to maintain fraud detection while reducing the false positives. What should the institution do?

   • ARoll back to a previous version of the model and stop updating it
   • BAdd more features to the model to improve its accuracy
   • CIncrease the model's decision threshold to approve more transactions
   • DImplement robust data validation and anomaly detection on the training data before it is used
   Reveal Solution Discussion Select an answer first
2. 27expert · hard

   A defense contractor develops an AI system for drone navigation. The system relies on GPS and visual data. An adversary could potentially spoof GPS signals or inject adversarial images to mislead the drone. The contractor must ensure the system is robust against these attacks while also maintaining real-time performance and avoiding false positives that could cause accidents. Which strategy is most effective?

   • ADisable the AI navigation and use manual remote control only.
   • BTrain the model exclusively on simulated adversarial examples to maximize robustness.
   • CUse a single, highly accurate sensor and rely on its precision.
   • DImplement sensor fusion with redundant, independent sources and use anomaly detection to flag inconsistent data.
   Reveal Solution Discussion Select an answer first
4. 28application · medium

   A logistics company plans to introduce autonomous delivery robots in a city. The robots will replace some human delivery workers. The company wants to minimize negative societal impact on employment while still gaining efficiency. What should they do?

   • ADeploy the robots immediately to maximize efficiency and cost savings.
   • BCancel the robot program entirely to protect jobs.
   • CKeep the robots but do not inform workers about the changes.
   • DOffer retraining programs for affected workers and phase in the robots gradually.
   Reveal Solution Discussion Select an answer first
5. 29expert · hard

   A social media company uses an AI system to moderate content. The system is trained to detect hate speech, but it has a higher false positive rate for posts written in certain dialects, leading to accusations of censorship. The company wants to reduce false positives while maintaining the system's ability to catch actual hate speech. The company has a limited budget and must choose a mitigation. Which approach best balances fairness and effectiveness?

   • AUse a human review team to double-check all flagged posts
   • BRemove the dialect feature from the model
   • CLower the model's sensitivity threshold to reduce false positives
   • DRetrain the model with more diverse examples of those dialects, including both hate speech and benign uses
   Reveal Solution Discussion Select an answer first
6. 30foundation · easy

   Which societal concern is raised when AI systems are used to make decisions that affect individuals, such as loan approvals or hiring, potentially reducing the ability of people to control their own lives?

   • AEmployment impact
   • BSecurity
   • CHuman autonomy
   • DInequality
   Reveal Solution Discussion Select an answer first