Why AI safety prevents misuse in Prompt Engineering / GenAI - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to define a simple AI safety check function.

Prompt Engineering / GenAI

def safety_check(input_text):
    if 'danger' in input_text:
        return [1]
    return 'safe'

Drag options to blanks, or click blank then click option'

A'block'

B'allow'

C'ignore'

D'warn'

Attempts:

3 left

💡 Hint

Common Mistakes

Returning 'allow' lets dangerous input pass.

Returning 'ignore' does not stop misuse.

The function should block inputs containing 'danger' to prevent misuse.

2fill in blank

medium

Complete the code to add a misuse prevention step before AI response.

Prompt Engineering / GenAI

def generate_response(user_input):
    if safety_check(user_input) == [1]:
        return 'Request blocked due to safety.'
    return 'Response generated'

Drag options to blanks, or click blank then click option'

A'warn'

B'allow'

C'block'

D'ignore'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'allow' lets unsafe input through.

Using 'ignore' does not block misuse.

If safety_check returns 'block', the response should be blocked to prevent misuse.

3fill in blank

hard

Fix the error in the misuse detection logic.

Prompt Engineering / GenAI

def misuse_detected(text):
    keywords = ['hack', 'steal', 'danger']
    for word in keywords:
        if word [1] text:
            return True
    return False

Drag options to blanks, or click blank then click option'

Ain

Bnot in

C==

D!=

Attempts:

3 left

💡 Hint

Common Mistakes

Using '==' compares whole strings, not membership.

Using 'not in' reverses the logic.

To check if a keyword is inside the text, use 'in' operator.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that flags unsafe inputs.

Prompt Engineering / GenAI

inputs = ['safe', 'hack', 'hello', 'steal']
flags = {item: item [1] keywords for item in inputs if item [2] keywords}

Drag options to blanks, or click blank then click option'

Ain

Bnot in

C==

D!=

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'not in' reverses the logic.

Using '==' only matches exact strings.

We check if item is in keywords to flag unsafe inputs correctly.

5fill in blank

hard

Fill all three blanks to filter and map unsafe words to warnings.

Prompt Engineering / GenAI

keywords = ['hack', 'steal', 'danger']
warnings = {word[1]: 'Warning!' for word in inputs if word [2] keywords and len(word) [3] 4}

Drag options to blanks, or click blank then click option'

A.upper()

Bin

C>

Dnot in

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'not in' excludes unsafe words.

Using '<' filters wrong length.

We convert words to uppercase, check if they are in keywords, and filter by length greater than 4.