Creating Malicious Software with AI GOAA Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A security operations center (SOC) uses an AI-based malware detector that has a low false positive rate but is known to be vulnerable to adversarial examples. An attacker is attempting to bypass the detector. Which technique would the attacker most likely use?

   • AAdd small perturbations to the malware's byte sequence that are designed to fool the AI model but do not affect its execution.
   • BEncrypt the malware with a new key each time to change its signature.
   • CRun the malware only in memory to avoid writing to disk.
   • DUse a packer to compress the malware and change its file structure.
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A media company wants to create a realistic deepfake of a historical figure for a documentary. The company has access to many photos and videos of the figure. Which technique would produce the most realistic results?

   • ATrain a GAN on the available images and videos to generate new footage of the figure.
   • BUse a simple morphing tool to blend between two photos of the figure.
   • CHire an actor to impersonate the figure and use face-swapping to replace the actor's face with the figure's.
   • DUse a video editing software to cut and paste existing footage of the figure.
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A malware researcher is building an AI system that automatically generates new malware variants to test an organization's defenses. The researcher must ensure the generated malware is functional and evades detection. Which approach best balances these requirements?

   • AUse a large language model to generate malware source code and assume it is functional because the model is trained on code.
   • BUse a reinforcement learning agent that directly generates binary files and is rewarded for evading the target detector.
   • CUse a generative model to create new source code, then compile and test each variant in a sandbox to verify functionality.
   • DUse a genetic algorithm to mutate existing malware binaries and select those that evade detection.
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which legal issue is most likely to arise from using AI to generate malicious software that evades detection?

   • AViolation of data protection regulations because the malware collects personal data.
   • BInfringement of patent rights for using a patented AI algorithm.
   • CViolation of computer fraud and abuse laws.
   • DBreach of copyright because the AI model was trained on open-source code.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A digital forensics investigator receives a video that allegedly shows a CEO making a fraudulent statement. The video has inconsistent blinking patterns and unnatural facial movements. Which deepfake creation technique is most likely to produce such artifacts?

   • AA face-swapping autoencoder that maps the target's face onto the source video frame-by-frame.
   • BA generative adversarial network (GAN) with a discriminator that penalizes unrealistic facial landmarks.
   • CA variational autoencoder (VAE) trained to reconstruct the CEO's face from a single image.
   • DA simple image filter that overlays the CEO's face onto the video without any machine learning.
   Reveal Solution Discussion Select an answer first