Knowledge Check – Module 2

What is a major data privacy concern when using AI tools that are not stored locally?

Local regulations are automatically applied to all data uploaded to external AI servers

Correct Incorrect Correct answer

Cloud-based AI is unable to recognize patterns in confidential datasets

Correct Incorrect Correct answer

Once data leaves the agency for the cloud, it is out of the organization's control

Correct Incorrect Correct answer

AI models will refuse to process data that has not been anonymized

Correct Incorrect Correct answer

An AI tool used for law enforcement is found to have a higher error rate for certain demographic groups. This is an example of:

A low-risk error that requires minimal human review

Correct Incorrect Correct answer

The AI successfully applying moral reasoning to complex cases

Correct Incorrect Correct answer

Algorithmic bias stemming from historical data

Correct Incorrect Correct answer

A prompt contradiction during the search process

Correct Incorrect Correct answer

Which mitigation strategy is appropriate for ‘Medium-Risk’ AI usage, such as drafting content or initial data analysis?

Liberal usage with almost no oversight or verification required

Correct Incorrect Correct answer

Immediate prohibition of the tool until all training data is public

Correct Incorrect Correct answer

Extensive legal and compliance reviews by multiple subject matter experts

Correct Incorrect Correct answer

A human review and fact-check of the generated content before use

Correct Incorrect Correct answer

When an AI system is asked to provide a list of legal precedents but includes a case that never existed, what has occurred?

A success in identifying patterns in judicial language

Correct Incorrect Correct answer

An appropriate use of low-risk general research

Correct Incorrect Correct answer

A sentence contradiction within the legal brief

Correct Incorrect Correct answer

A factual contradiction or random hallucination

Correct Incorrect Correct answer

A public sector employee notices that an AI-generated response provides a factual claim in the first paragraph but then refutes that same claim in the third paragraph. Which type of AI hallucination is this?

Sentence contradiction

Correct Incorrect Correct answer

Prompt contradiction

Correct Incorrect Correct answer

Factual contradiction

Correct Incorrect Correct answer

Random hallucination

Correct Incorrect Correct answer

Why is the ‘confidence of presentation’ in an AI’s output a potential risk for public sector employees?

It can mask inaccuracies, such as fake citations or invented statistics

Correct Incorrect Correct answer

It indicates that the AI has verified the information against local regulations

Correct Incorrect Correct answer

It ensures that the training data used was high-quality and complete

Correct Incorrect Correct answer

It prevents the AI from producing biased outcomes in sensitive areas

Correct Incorrect Correct answer

In the context of AI training data, what is the primary consequence of the ‘garbage in, garbage out’ principle?

Poor or incomplete data leads to unreliable and potentially unfair results

Correct Incorrect Correct answer

It causes the AI to lose the ability to process large amounts of data quickly

Correct Incorrect Correct answer

It forces the AI to operate only on low-risk tasks like document formatting

Correct Incorrect Correct answer

The AI system will automatically flag data that contains historical discrimination

Correct Incorrect Correct answer

Cart