CompTIA Security+ SY0-801 Subdomain 2.6 Practice Test #1

This practice test is based on Subdomain 2.6 (Summarize threats and vulnerabilities associated with artificial intelligence (AI) usage) from the CompTIA Security+ SY0-801 objectives.

This beginner-level practice test is inspired by the CompTIA Security+ (SY0-801) exam and is designed to help you reinforce key cybersecurity concepts.

Browse CompTIA Security+ SY0-801 practice tests by domain and subdomain.

Practice test disclaimer

These questions are not official exam questions, nor are they brain dumps. They reflect topics and scenarios relevant to the Security+ certification. Use them to test your knowledge and identify areas for improvement.

CompTIA and Security+ are registered trademarks of CompTIA. This content is not affiliated with or endorsed by CompTIA.

Security+ SY0-801 Objective 2.6 Practice Test 1
10 questions • Single best answer
Question 1
A support assistant summarizes incoming mail. One customer's message contains directions addressed to the assistant that have nothing to do with the support case. What trust boundary should the application enforce?
    Question 2
    A screening model consistently rejects qualified applicants from one group. Test records show the group's training examples are sparse and historically skewed. What risk best fits these observations?
      Question 3
      A legal research assistant cites a plausible judgment with a court, date, and docket number. The court's records confirm that the citation does not exist. What should the reviewer identify?
        Question 4
        Analysts discover a batch of deliberately mislabeled examples in the training data for a fraud model. Its behavior changes after retraining on that batch. Which threat best fits the timing and evidence?
          Question 5
          An internal assistant answers employee questions from stored documents. It quotes salary details from a folder the requesting employee cannot open. Which correction most directly addresses the exposure?
            Question 6
            A claims model predicts accurately on held-out cases. Reviewers cannot explain what drove an individual rejection. Which property is missing?
              Question 7
              A team wants to run AI-generated programs. Its proposed execution account has production administrator rights. Which change best limits damage from a faulty program?
                Question 8
                An employee pastes a customer file into a public chatbot to summarize it. The summary contains no names, but the provider already received the original input. Where did the privacy exposure occur?
                  Question 9
                  A security classifier performs well on normal inputs. During testing, a slightly altered input is misclassified by the unchanged model while a human still recognizes the same content. Which risk does this test illustrate?
                    Question 10
                    Users frequently accept an AI assistant's recommendations. A manager wants to detect persuasive but inaccurate answers. What else should the team measure?
                      Scroll to Top