Knowing when NOT to use AI in AI for Everyone - Time & Space Complexity

Time Complexity: Knowing when NOT to use AI

O(n)

Understanding Time Complexity

We want to understand how the cost of using AI changes as the problem size grows.

Specifically, when does using AI become too costly or inefficient?

Scenario Under Consideration

Analyze the time complexity of deciding when NOT to use AI.

function shouldUseAI(task) {
  if (task.isSimple) {
    return false;
  }
  if (task.requiresHumanJudgment) {
    return false;
  }
  if (task.dataIsInsufficient) {
    return false;
  }
  return true;
}
    

This code checks simple conditions to decide if AI should be used or not.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: A few simple condition checks.
• How many times: Each condition is checked once per task.

How Execution Grows With Input

The number of checks stays the same no matter how big the task is.

Input Size (n)	Approx. Operations
10	3 checks per task x 10 = 30 checks
100	3 checks per task x 100 = 300 checks
1000	3 checks per task x 1000 = 3000 checks

Pattern observation: The work grows directly with the number of tasks, but each task only needs a fixed small number of checks.

Final Time Complexity

Time Complexity: O(n)

This means the time to decide grows in a straight line with the number of tasks.

Common Mistake

[X] Wrong: "Checking if AI should be used takes a lot of time because AI itself is complex."

[OK] Correct: The decision to use AI is usually a simple check done once per task, not the AI processing itself.

Interview Connect

Understanding when not to use AI shows you can think clearly about costs and benefits, a skill valuable in many real-world situations.

Self-Check

"What if the decision to use AI involved checking many complex conditions or data points? How would the time complexity change?"

Practice

1. Which situation is NOT a good reason to use AI?

easy

A. When automating repetitive and simple tasks

B. When you need to analyze large amounts of data quickly

C. When the task requires deep human empathy and understanding

D. When you want to speed up decision-making with clear rules

Solution

1. Step 1: Understand AI strengths

   AI excels at processing large data and automating repetitive tasks quickly.

2. Step 2: Identify human-only tasks

   Tasks needing deep empathy and understanding are better handled by humans.

3. Final Answer:

   When the task requires deep human empathy and understanding -> Option C

4. Quick Check:

   Human empathy ≠ AI strength [OK]

Hint: Choose human skills for empathy, not AI [OK]

Common Mistakes:

• Thinking AI can replace human emotions
• Assuming AI is best for all tasks
• Ignoring privacy concerns

2. Which of the following is the correct reason NOT to use AI?

easy

A. To protect sensitive personal data from being exposed

B. To speed up calculations on large datasets

C. To automate routine email sorting

D. To generate quick summaries of documents

Solution

1. Step 1: Identify privacy concerns

   Using AI on sensitive data risks exposure or misuse of private information.

2. Step 2: Compare with other options

   Other options describe good uses of AI, not reasons to avoid it.

3. Final Answer:

   To protect sensitive personal data from being exposed -> Option A

4. Quick Check:

   Privacy risk = avoid AI [OK]

Hint: Avoid AI when privacy is at risk [OK]

Common Mistakes:

• Confusing AI benefits with privacy risks
• Choosing options that describe good AI uses
• Ignoring data sensitivity

3. Consider this scenario: You have a confidential medical record that needs review. Which action is safest?

medium

A. Use AI to analyze the record for diagnosis suggestions

B. Have a trained doctor review the record manually

C. Upload the record to a public AI service for quick results

D. Share the record with multiple AI tools to compare outputs

Solution

1. Step 1: Assess privacy and accuracy needs

   Medical records are sensitive and require expert human review for accuracy and confidentiality.

2. Step 2: Evaluate options for safety

   Using AI publicly risks data leaks; human experts protect privacy and provide reliable judgment.

3. Final Answer:

   Have a trained doctor review the record manually -> Option B

4. Quick Check:

   Privacy + expertise = human review [OK]

Hint: Choose human experts for sensitive medical data [OK]

Common Mistakes:

• Assuming AI is always faster and safer
• Ignoring data confidentiality
• Trusting public AI services with private info

4. You wrote a program that sends sensitive user data to an AI service without encryption. What is the main problem?

medium

A. The AI service will refuse to process unencrypted data

B. AI cannot analyze unencrypted data correctly

C. The program will run slower because of missing encryption

D. User data privacy is at risk due to insecure transmission

Solution

1. Step 1: Understand data transmission risks

   Sending sensitive data without encryption exposes it to interception and misuse.

2. Step 2: Analyze options

   AI services usually accept data; speed and analysis correctness are not affected by encryption presence.

3. Final Answer:

   User data privacy is at risk due to insecure transmission -> Option D

4. Quick Check:

   Unencrypted data = privacy risk [OK]

Hint: Encrypt sensitive data before sending to AI [OK]

Common Mistakes:

• Thinking AI rejects unencrypted data
• Confusing speed with security
• Assuming AI needs encrypted data to work

5. You want to automate customer support but some questions need personal judgment and empathy. What is the best approach?

hard

A. Use AI for simple questions and humans for empathetic ones

B. Avoid AI completely to keep all responses personal

C. Use AI for all questions to save time

D. Train AI to fully replace human empathy

Solution

1. Step 1: Identify task types

   Simple, repetitive questions can be handled by AI; empathetic questions need humans.

2. Step 2: Combine AI and human strengths

   Using AI for routine tasks and humans for empathy balances efficiency and care.

3. Final Answer:

   Use AI for simple questions and humans for empathetic ones -> Option A

4. Quick Check:

   Mix AI + human empathy = best approach [OK]

Hint: Split tasks: AI for simple, humans for empathy [OK]

Common Mistakes:

• Relying solely on AI for empathy
• Ignoring efficiency gains from AI
• Avoiding AI even for simple tasks