Audio, Image, and Video Deepfakes GOAA Practice Questions (Page 8)

Questions 36–40

1. 36application · medium

   A news organization receives a leaked video that appears to show a politician accepting a bribe. The video is grainy and low-resolution. Before airing it, the news team wants to verify its authenticity. Which step is most appropriate?

   • AContact the politician's office for a statement before airing.
   • BBroadcast the video with a disclaimer that it may be a deepfake, citing the low quality.
   • CCheck the video's metadata for the original recording device and date.
   • DEnhance the video using super-resolution techniques and then look for deepfake artifacts.
   Reveal Solution Discussion Select an answer first
2. 37application · medium

   A legislative body is drafting a law to address malicious deepfakes. Which provision would best balance free speech and protection from harm?

   • AAllow deepfakes only for government and military purposes.
   • BRequire all deepfakes to be labeled, regardless of context.
   • CCriminalize deepfakes created with the intent to deceive and cause harm, while exempting satire and parody.
   • DBan all deepfake creation and distribution outright.
   Reveal Solution Discussion Select an answer first
4. 38expert · hard

   A social media company is deploying a deepfake detection system. They must balance user privacy with the need to scan uploaded content. Which approach best addresses both concerns?

   • AScan all uploads on the server using a powerful model, and store the results for future analysis.
   • BPerform on-device detection using a lightweight model that runs on the user's device, sending only flagged content to the server.
   • COnly scan content that has been reported by users.
   • DUse a third-party service to analyze all uploads, relying on their privacy policy.
   Reveal Solution Discussion Select an answer first
5. 39application · medium

   A government agency is drafting guidelines for the use of deepfake technology in official training simulations. Which ethical principle should be the primary consideration?

   • AObtain informed consent from any individuals whose likeness is used, even in fictional scenarios.
   • BEnsure the simulations are as realistic as possible to maximize training effectiveness.
   • CClassify all deepfake training materials as top secret to prevent public awareness.
   • DAvoid using any real person's likeness and only use fully synthetic characters.
   Reveal Solution Discussion Select an answer first
6. 40foundation · easy

   Which generative model is specifically designed to create realistic facial images by pitting a generator against a discriminator?

   • AGenerative adversarial network (GAN)
   • BConvolutional neural network (CNN)
   • CRecurrent neural network (RNN)
   • DSupport vector machine (SVM)
   Reveal Solution Discussion Select an answer first