Microsoft AI-103 Dumps
Microsoft AI-103 Practice Test 77
Question 760
During a production incident, a data engineer must accept a change for video object identification. The issue to be addressed is: A traffic analysis team must identify vehicles across a series of frames. Which concrete verification action demonstrates that the relevant behavior works?
Explanation
Correct Answer
To verify this change, the industrial maintenance team should verify detected vehicle movement across annotated frames. Operators still need the implementation to use a temporally aware visual understanding workflow and preserve object evidence and the preventive requirement to define confidence thresholds and avoid asserting identity without evidence.
Incorrect Answers
In the industrial maintenance team acceptance review, trying to infer cross-frame identity solely from image filenames does not test the exact video object identification condition where a traffic analysis team must identify vehicles across a series of frames; the relevant test is to verify detected vehicle movement across annotated frames. Acceptance should instead demonstrate that the service can verify detected vehicle movement across annotated frames, using the described failure as the test criterion.
The action to use a temporally aware visual understanding workflow and preserve object evidence can contribute to video object identification as a implementation step; here, a traffic analysis team must identify vehicles across a series of frames. In the industrial maintenance team acceptance review, trying to use a temporally aware visual understanding workflow and preserve object evidence does not test the exact video object identification condition where a traffic analysis team must identify vehicles across a series of frames; the relevant test is to verify detected vehicle movement across annotated frames.
For content understanding pro-mode, it can be sensible to run representative multi-step examples and inspect intermediate outputs if an inspection task spans multiple related media interpretation subtasks. Without the actual change to verify detected vehicle movement across annotated frames, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.
Question 761
At a transport authority, the service administrator is preparing a architecture migration. The concrete requirement is: An image generation tool produces assets outside the organization's approved visual policy. Which design action BEST addresses the stated need?
Explanation
Correct Answer
The proposed implementation is to apply appropriate visual safety screening and prompt restrictions before display; this matches the operational requirement that an image generation tool produces assets outside the organization's approved visual policy. The team can test both permitted and disallowed visual content cases. The team should also maintain a visual policy test set and manual escalation path.
Incorrect Answers
Choosing to assume text-only chat moderation guarantees generated image safety cannot directly replace the specified action to apply appropriate visual safety screening and prompt restrictions before display when an image generation tool produces assets outside the organization's approved visual policy. The actual implementation must apply appropriate visual safety screening and prompt restrictions before display, because the stated service condition is that an image generation tool produces assets outside the organization's approved visual policy.
The action to test both permitted and disallowed visual content cases can contribute to visual content policy as a targeted acceptance check step; here, an image generation tool produces assets outside the organization's approved visual policy. Choosing to test both permitted and disallowed visual content cases cannot directly replace the specified action to apply appropriate visual safety screening and prompt restrictions before display when an image generation tool produces assets outside the organization's approved visual policy.
Choosing to apply supported provenance labeling or watermark controls where required cannot directly replace the specified action to apply appropriate visual safety screening and prompt restrictions before display when an image generation tool produces assets outside the organization's approved visual policy. The relevant corrective or preventive action here remains to apply appropriate visual safety screening and prompt restrictions before display, with its own validation criteria.
Question 762
After a operational readiness review at a large online retailer, the team observes this failure: An image generation tool produces assets outside the organization's approved visual policy. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?
Explanation
Correct Answer
This design mistake is to assume text-only chat moderation guarantees generated image safety; it conflicts with the actual visual content policy behavior needed when an image generation tool produces assets outside the organization's approved visual policy. The large online retailer should instead apply appropriate visual safety screening and prompt restrictions before display, followed by a check to test both permitted and disallowed visual content cases.
Incorrect Answers
The alternative to test both permitted and disallowed visual content cases is a targeted acceptance check step rather than the faulty assumption being diagnosed. This large online retailer failure-analysis alternative of test both permitted and disallowed visual content cases does not diagnose the particular visual content policy defect where an image generation tool produces assets outside the organization's approved visual policy; the keyed mistake is assume text-only chat moderation guarantees generated image safety.
The option to assume source footage being on the web grants permission to clone the person would normally serve a different video rights and consent requirement, where a team wants to generate a likeness of a real speaker without permission. The alternative may be useful in its own context, but it does not directly supply assume text-only chat moderation guarantees generated image safety for the stated decision.
The alternative to apply appropriate visual safety screening and prompt restrictions before display is a implementation step rather than the faulty assumption being diagnosed. This large online retailer failure-analysis alternative of apply appropriate visual safety screening and prompt restrictions before display does not diagnose the particular visual content policy defect where an image generation tool produces assets outside the organization's approved visual policy; the keyed mistake is assume text-only chat moderation guarantees generated image safety.
Question 763
The public library network is responding to a regional deployment. A site reliability engineer discovers this issue: An image generation tool produces assets outside the organization's approved visual policy. What corrective step should the team take to restore the intended functionality most directly?
Explanation
Correct Answer
The direct repair is to block unsafe assets and revise the visual safety configuration. Once the condition is corrected, the public library network should test both permitted and disallowed visual content cases. To reduce recurrence, it should also maintain a visual policy test set and manual escalation path; those steps provide different recovery and prevention functions.
Incorrect Answers
The action to test both permitted and disallowed visual content cases can contribute to visual content policy as a targeted acceptance check step; here, an image generation tool produces assets outside the organization's approved visual policy. During public library network remediation, the proposed step to test both permitted and disallowed visual content cases is not the direct restoration action for visual content policy when an image generation tool produces assets outside the organization's approved visual policy; the keyed repair is block unsafe assets and revise the visual safety configuration.
During public library network remediation, the proposed step to assume text-only chat moderation guarantees generated image safety is not the direct restoration action for visual content policy when an image generation tool produces assets outside the organization's approved visual policy; the keyed repair is block unsafe assets and revise the visual safety configuration. After remediation the team must test both permitted and disallowed visual content cases; the action in this option does not itself establish that target recovery outcome.
During public library network remediation, the proposed step to add approved disclosure metadata or watermark handling before release is not the direct restoration action for visual content policy when an image generation tool produces assets outside the organization's approved visual policy; the keyed repair is block unsafe assets and revise the visual safety configuration. If the team chose it instead, the targeted check to test both permitted and disallowed visual content cases could still fail because the underlying required action is to block unsafe assets and revise the visual safety configuration.
Question 764
A security engineer at a software consultancy wants the following issue to be prevented in later deployments: An image generation tool produces assets outside the organization's approved visual policy. Which enduring control or engineering guardrail BEST prevents recurrence?
Explanation
Correct Answer
Standardizing the boundary to maintain a visual policy test set and manual escalation path is preventive because the same visual content policy failure could otherwise return: an image generation tool produces assets outside the organization's approved visual policy. For this software consultancy, the service should also apply appropriate visual safety screening and prompt restrictions before display, and tests should test both permitted and disallowed visual content cases so that the ongoing requirement is measurable.
Incorrect Answers
The action to block unsafe assets and revise the visual safety configuration can contribute to visual content policy as a corrective remediation step; here, an image generation tool produces assets outside the organization's approved visual policy. For prevention at the software consultancy, the measure to block unsafe assets and revise the visual safety configuration is weaker or aimed at another issue than the visual content policy guardrail to maintain a visual policy test set and manual escalation path needed when an image generation tool produces assets outside the organization's approved visual policy.
For prevention at the software consultancy, the measure to assume text-only chat moderation guarantees generated image safety is weaker or aimed at another issue than the visual content policy guardrail to maintain a visual policy test set and manual escalation path needed when an image generation tool produces assets outside the organization's approved visual policy. The lasting guardrail for this specific threat is to maintain a visual policy test set and manual escalation path, not simply to complete an adjacent service task.
The action to include provenance checks in the media publication workflow addresses generated media watermark policy rather than the decisive visual content policy condition. The software consultancy should not infer that fixing one adjacent AI service validates its unrelated workflow; the requested action is to maintain a visual policy test set and manual escalation path.
Question 765
During a operational readiness review, a agent designer must accept a change for visual content policy. The issue to be addressed is: An image generation tool produces assets outside the organization's approved visual policy. Which concrete verification action demonstrates that the relevant behavior works?
Explanation
Correct Answer
The targeted acceptance test is to test both permitted and disallowed visual content cases. The design should first apply appropriate visual safety screening and prompt restrictions before display. A robust regional hospital deployment also needs to maintain a visual policy test set and manual escalation path; passing this check provides relevant evidence but does not certify every possible input or security risk.
Incorrect Answers
In the regional hospital acceptance review, trying to assume text-only chat moderation guarantees generated image safety does not test the exact visual content policy condition where an image generation tool produces assets outside the organization's approved visual policy; the relevant test is to test both permitted and disallowed visual content cases. Acceptance should instead demonstrate that the service can test both permitted and disallowed visual content cases, using the described failure as the test criterion.
The action to block unsafe assets and revise the visual safety configuration can contribute to visual content policy as a corrective remediation step; here, an image generation tool produces assets outside the organization's approved visual policy. In the regional hospital acceptance review, trying to block unsafe assets and revise the visual safety configuration does not test the exact visual content policy condition where an image generation tool produces assets outside the organization's approved visual policy; the relevant test is to test both permitted and disallowed visual content cases.
The action to maintain a visual policy test set and manual escalation path can contribute to visual content policy as a ongoing preventive rule step; here, an image generation tool produces assets outside the organization's approved visual policy. In the regional hospital acceptance review, trying to maintain a visual policy test set and manual escalation path does not test the exact visual content policy condition where an image generation tool produces assets outside the organization's approved visual policy; the relevant test is to test both permitted and disallowed visual content cases.
The action to apply appropriate visual safety screening and prompt restrictions before display can contribute to visual content policy as a implementation step; here, an image generation tool produces assets outside the organization's approved visual policy. In the regional hospital acceptance review, trying to apply appropriate visual safety screening and prompt restrictions before display does not test the exact visual content policy condition where an image generation tool produces assets outside the organization's approved visual policy; the relevant test is to test both permitted and disallowed visual content cases.
The option to generate adversarial logo prompts and confirm review or rejection would normally serve a different brand misuse detection requirement, where a marketing generator sometimes produces unapproved partner logos. The alternative may be useful in its own context, but it does not directly supply test both permitted and disallowed visual content cases for the stated decision.
Question 766
At a industrial maintenance team, the responsible-AI reviewer is preparing a cost review. The concrete requirement is: A photographed document contains tiny text telling the agent to execute a secret-export tool. Which design action BEST addresses the stated need?
Explanation
Correct Answer
Choosing to treat embedded image text as untrusted task data rather than higher-level instructions directly targets the described image embedded prompt injection gap: a photographed document contains tiny text telling the agent to execute a secret-export tool. As a lasting boundary, they also need to test cross-modal prompt-injection and enforce tool permission checks, because an implementation by itself does not prove safe operation.
Incorrect Answers
The action to show an adversarial document image and verify no unauthorized action can contribute to image embedded prompt injection as a targeted acceptance check step; here, a photographed document contains tiny text telling the agent to execute a secret-export tool. Choosing to show an adversarial document image and verify no unauthorized action cannot directly replace the specified action to treat embedded image text as untrusted task data rather than higher-level instructions when a photographed document contains tiny text telling the agent to execute a secret-export tool.
Choosing to execute commands extracted from the image without checking authorization cannot directly replace the specified action to treat embedded image text as untrusted task data rather than higher-level instructions when a photographed document contains tiny text telling the agent to execute a secret-export tool. The actual implementation must treat embedded image text as untrusted task data rather than higher-level instructions, because the stated service condition is that a photographed document contains tiny text telling the agent to execute a secret-export tool.
Choosing to require appropriate rights and consent before producing identifiable synthetic media would direct effort toward video rights and consent, whose characteristic issue is that a team wants to generate a likeness of a real speaker without permission. The relevant corrective or preventive action here remains to treat embedded image text as untrusted task data rather than higher-level instructions, with its own validation criteria.
Question 767
After a reliability exercise at a research university, the team observes this failure: A photographed document contains tiny text telling the agent to execute a secret-export tool. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?
Explanation
Correct Answer
The faulty assumption is execute commands extracted from the image without checking authorization. A functioning alternative would be to treat embedded image text as untrusted task data rather than higher-level instructions, and the relevant validation would be to show an adversarial document image and verify no unauthorized action.
Incorrect Answers
The alternative to treat embedded image text as untrusted task data rather than higher-level instructions is a implementation step rather than the faulty assumption being diagnosed. This research university failure-analysis alternative of treat embedded image text as untrusted task data rather than higher-level instructions does not diagnose the particular image embedded prompt injection defect where a photographed document contains tiny text telling the agent to execute a secret-export tool; the keyed mistake is execute commands extracted from the image without checking authorization.
The alternative to show an adversarial document image and verify no unauthorized action is a targeted acceptance check step rather than the faulty assumption being diagnosed. This research university failure-analysis alternative of show an adversarial document image and verify no unauthorized action does not diagnose the particular image embedded prompt injection defect where a photographed document contains tiny text telling the agent to execute a secret-export tool; the keyed mistake is execute commands extracted from the image without checking authorization.
This research university failure-analysis alternative of treat any familiar logo in generated content as legally permitted does not diagnose the particular image embedded prompt injection defect where a photographed document contains tiny text telling the agent to execute a secret-export tool; the keyed mistake is execute commands extracted from the image without checking authorization. The alternative may be useful in its own context, but it does not directly supply execute commands extracted from the image without checking authorization for the stated decision.
Question 768
The education technology company is responding to a source update. A machine-learning developer discovers this issue: A photographed document contains tiny text telling the agent to execute a secret-export tool. What corrective step should the team take to restore the intended functionality most directly?
Explanation
Correct Answer
The appropriate remediation is to remove the injected text from the instruction channel and retry interpretation, because the failure concerns image embedded prompt injection: a photographed document contains tiny text telling the agent to execute a secret-export tool. A meaningful verification should show an adversarial document image and verify no unauthorized action, and the durable operating rule should test cross-modal prompt-injection and enforce tool permission checks after the immediate fix.
Incorrect Answers
This approach relates to brand misuse detection: a marketing generator sometimes produces unapproved partner logos. If the team chose it instead, the targeted check to show an adversarial document image and verify no unauthorized action could still fail because the underlying required action is to remove the injected text from the instruction channel and retry interpretation.
The action to show an adversarial document image and verify no unauthorized action can contribute to image embedded prompt injection as a targeted acceptance check step; here, a photographed document contains tiny text telling the agent to execute a secret-export tool. During education technology company remediation, the proposed step to show an adversarial document image and verify no unauthorized action is not the direct restoration action for image embedded prompt injection when a photographed document contains tiny text telling the agent to execute a secret-export tool; the keyed repair is remove the injected text from the instruction channel and retry interpretation.
During education technology company remediation, the proposed step to execute commands extracted from the image without checking authorization is not the direct restoration action for image embedded prompt injection when a photographed document contains tiny text telling the agent to execute a secret-export tool; the keyed repair is remove the injected text from the instruction channel and retry interpretation. After remediation the team must show an adversarial document image and verify no unauthorized action; the action in this option does not itself establish that target recovery outcome.
Question 769
A AI application owner at a pharmaceutical research group wants the following issue to be prevented in later deployments: A photographed document contains tiny text telling the agent to execute a secret-export tool. Which enduring control or engineering guardrail BEST prevents recurrence?
Explanation
Correct Answer
The sustained prevention measure is to test cross-modal prompt-injection and enforce tool permission checks. The implementation can still use the separate action to treat embedded image text as untrusted task data rather than higher-level instructions, but a guardrail constrains future changes. The pharmaceutical research group can measure whether the rule works by arranging to show an adversarial document image and verify no unauthorized action.
Incorrect Answers
The action to maintain brand rules and rights checks for output assets addresses brand misuse detection rather than the decisive image embedded prompt injection condition. The pharmaceutical research group should not infer that fixing one adjacent AI service validates its unrelated workflow; the requested action is to test cross-modal prompt-injection and enforce tool permission checks.
For prevention at the pharmaceutical research group, the measure to execute commands extracted from the image without checking authorization is weaker or aimed at another issue than the image embedded prompt injection guardrail to test cross-modal prompt-injection and enforce tool permission checks needed when a photographed document contains tiny text telling the agent to execute a secret-export tool. The lasting guardrail for this specific threat is to test cross-modal prompt-injection and enforce tool permission checks, not simply to complete an adjacent service task.
The action to remove the injected text from the instruction channel and retry interpretation can contribute to image embedded prompt injection as a corrective remediation step; here, a photographed document contains tiny text telling the agent to execute a secret-export tool. For prevention at the pharmaceutical research group, the measure to remove the injected text from the instruction channel and retry interpretation is weaker or aimed at another issue than the image embedded prompt injection guardrail to test cross-modal prompt-injection and enforce tool permission checks needed when a photographed document contains tiny text telling the agent to execute a secret-export tool.
Question 770
During a reliability exercise, a quality assurance lead must accept a change for image embedded prompt injection. The issue to be addressed is: A photographed document contains tiny text telling the agent to execute a secret-export tool. Which concrete verification action demonstrates that the relevant behavior works?
Explanation
Correct Answer
To verify this change, the travel support provider should show an adversarial document image and verify no unauthorized action. Operators still need the implementation to treat embedded image text as untrusted task data rather than higher-level instructions and the preventive requirement to test cross-modal prompt-injection and enforce tool permission checks.
Incorrect Answers
The action to test cross-modal prompt-injection and enforce tool permission checks can contribute to image embedded prompt injection as a ongoing preventive rule step; here, a photographed document contains tiny text telling the agent to execute a secret-export tool. In the travel support provider acceptance review, trying to test cross-modal prompt-injection and enforce tool permission checks does not test the exact image embedded prompt injection condition where a photographed document contains tiny text telling the agent to execute a secret-export tool; the relevant test is to show an adversarial document image and verify no unauthorized action.
In the travel support provider acceptance review, trying to execute commands extracted from the image without checking authorization does not test the exact image embedded prompt injection condition where a photographed document contains tiny text telling the agent to execute a secret-export tool; the relevant test is to show an adversarial document image and verify no unauthorized action. Acceptance should instead demonstrate that the service can show an adversarial document image and verify no unauthorized action, using the described failure as the test criterion.
Choosing to submit benign photos with adversarial captions and inspect the response would direct effort toward unsafe image-to-text output, whose characteristic issue is that a vision assistant responds to benign photos with unsafe instructions unrelated to the image. The relevant corrective or preventive action here remains to show an adversarial document image and verify no unauthorized action, with its own validation criteria.
The action to treat embedded image text as untrusted task data rather than higher-level instructions can contribute to image embedded prompt injection as a implementation step; here, a photographed document contains tiny text telling the agent to execute a secret-export tool. In the travel support provider acceptance review, trying to treat embedded image text as untrusted task data rather than higher-level instructions does not test the exact image embedded prompt injection condition where a photographed document contains tiny text telling the agent to execute a secret-export tool; the relevant test is to show an adversarial document image and verify no unauthorized action.
Developing AI Apps and Agents on Azure (AI-103) Practice Tests
- Microsoft AI-103 Practice Test 75 - 10 Questions (Last Reviewed 04-Nov-2026)
- Microsoft AI-103 Practice Test 76 - 10 Questions (Last Reviewed 03-Nov-2026)
- Microsoft AI-103 Practice Test 77 - 10 Questions (Last Reviewed 02-Nov-2026)
- Microsoft AI-103 Practice Test 78 - 10 Questions (Last Reviewed 01-Nov-2026)
- Microsoft AI-103 Practice Test 79 - 10 Questions (Last Reviewed 31-Oct-2026)