Bias in generative models in Prompt Engineering / GenAI - Model Metrics & Evaluation

Metrics & Evaluation - Bias in generative models

Which metric matters for Bias in generative models and WHY

Bias in generative models means the model creates outputs that unfairly favor or harm certain groups. To measure this, we use fairness metrics like demographic parity or equal opportunity. These metrics check if the model treats different groups equally in its outputs. We also look at diversity metrics to see if the model generates a wide range of ideas or just repeats stereotypes. These metrics matter because they help us find and fix unfair or harmful patterns in the model's creations.

Confusion matrix or equivalent visualization

For bias, a confusion matrix is less direct. Instead, we use group-wise outcome tables. For example, if a model generates job recommendations, we count how many times each group (e.g., men, women) gets recommended for high-paying jobs.

Group          | Recommended | Not Recommended | Total
-------------------------------------------------------
Men            | 80          | 20              | 100
Women          | 50          | 50              | 100

Here, men get recommended 80% of the time, women only 50%. This shows bias favoring men.

Precision vs Recall tradeoff (or equivalent) with concrete examples

In bias evaluation, the tradeoff is between fairness and utility. For example, a generative model might produce very accurate text but repeat harmful stereotypes (high utility, low fairness). Or it might avoid stereotypes but produce less relevant content (high fairness, lower utility).

Example: A chatbot that answers questions. If it tries to be fair by avoiding certain topics, it might miss some correct answers (lower recall). If it answers everything without filtering, it might produce biased or offensive content (low fairness).

What "good" vs "bad" metric values look like for this use case

Good: Similar recommendation rates across groups (e.g., men 75%, women 73%), showing fairness. Diverse outputs covering many perspectives. Low bias scores in fairness metrics.

Bad: Large gaps in group outcomes (e.g., men 90%, women 40%), showing bias. Repetitive or stereotyped outputs. High bias scores indicating unfair treatment.

Metrics pitfalls

• Ignoring context: Some bias metrics miss subtle harms or cultural differences.
• Data leakage: If training data is biased, metrics may falsely show good fairness.
• Overfitting fairness: Fixing bias on one metric may cause worse bias elsewhere.
• Accuracy paradox: A model can be accurate but still biased.

Self-check question

Your generative model creates text with 98% accuracy on a test set but shows 30% fewer positive outcomes for one group compared to another. Is it good for production? Why or why not?

Answer: No, because despite high accuracy, the model treats groups unfairly. This bias can cause harm or legal issues. You should improve fairness before production.

Key Result

Fairness and diversity metrics are key to detect and reduce bias in generative models, ensuring outputs treat all groups fairly.

Practice

1. What is the main cause of bias in generative AI models?

easy

A. The speed of the computer

B. The programming language used

C. The data used to train the model

D. The color of the user interface

Solution

1. Step 1: Understand what bias means in generative models

   Bias means the model gives unfair or unbalanced results.

2. Step 2: Identify the source of bias

   Bias mainly comes from the data used to train the model, as it reflects existing patterns or prejudices.

3. Final Answer:

   The data used to train the model -> Option C

4. Quick Check:

   Bias source = training data [OK]

Hint: Bias mostly comes from training data, not code or hardware [OK]

Common Mistakes:

• Thinking bias comes from programming language
• Blaming hardware speed for bias
• Confusing UI design with bias

2. Which of the following is the correct way to describe bias in generative models?

easy

A. Bias means the model produces unfair or unbalanced outputs

B. Bias means the model always predicts correctly

C. Bias means the model runs faster on some computers

D. Bias means the model uses more memory

Solution

1. Step 1: Define bias in the context of generative models

   Bias refers to unfair or unbalanced outputs, not performance or resource use.

2. Step 2: Match the correct description

   Bias means the model produces unfair or unbalanced outputs correctly states bias as unfair or unbalanced outputs.

3. Final Answer:

   Bias means the model produces unfair or unbalanced outputs -> Option A

4. Quick Check:

   Bias = unfair outputs [OK]

Hint: Bias is about fairness in output, not speed or memory [OK]

Common Mistakes:

• Confusing bias with model accuracy
• Mixing bias with hardware performance
• Thinking bias relates to memory use

3. Consider a generative model trained on text data mostly from one culture. What is likely to happen when it generates stories about other cultures?

medium

A. It may produce biased or stereotyped stories about other cultures

B. It will generate perfectly balanced stories about all cultures

C. It will refuse to generate any story about other cultures

D. It will generate stories faster for other cultures

Solution

1. Step 1: Understand training data influence

   The model learns patterns from its training data, so if data is mostly from one culture, it lacks diversity.

2. Step 2: Predict output behavior

   When asked about other cultures, the model may produce biased or stereotyped stories due to limited or skewed data.

3. Final Answer:

   It may produce biased or stereotyped stories about other cultures -> Option A

4. Quick Check:

   Limited data causes biased outputs [OK]

Hint: Limited data diversity causes biased outputs [OK]

Common Mistakes:

• Assuming model is unbiased regardless of data
• Thinking model refuses to generate unknown topics
• Confusing speed with bias

4. You notice your generative model outputs biased text favoring one gender. Which step can help fix this issue?

medium

A. Use a smaller batch size during training

B. Increase the model's learning rate

C. Reduce the number of training epochs

D. Add more balanced and diverse training data

Solution

1. Step 1: Identify cause of bias

   Bias often comes from unbalanced training data that favors one group.

2. Step 2: Choose corrective action

   Adding more balanced and diverse data helps the model learn fairer patterns and reduce bias.

3. Final Answer:

   Add more balanced and diverse training data -> Option D

4. Quick Check:

   Balanced data reduces bias [OK]

Hint: Fix bias by improving training data diversity [OK]

Common Mistakes:

• Changing learning rate without addressing data
• Adjusting batch size unrelated to bias
• Reducing epochs without fixing data

5. A company wants to reduce bias in its generative model that creates job descriptions. Which combined approach is best to improve fairness?

hard

A. Remove all rare words from the training data

B. Use diverse training data and add fairness constraints during model training

C. Train the model faster with fewer epochs

D. Only increase the model size without changing data

Solution

1. Step 1: Understand bias reduction methods

   Bias can be reduced by improving data diversity and applying fairness rules during training.

2. Step 2: Evaluate options

   Use diverse training data and add fairness constraints during model training combines better data and fairness constraints, which is more effective than just changing model size or training speed.

3. Final Answer:

   Use diverse training data and add fairness constraints during model training -> Option B

4. Quick Check:

   Data + fairness constraints = less bias [OK]

Hint: Combine diverse data with fairness rules for best bias fix [OK]

Common Mistakes:

• Thinking bigger model alone fixes bias
• Speeding training reduces bias (it doesn't)
• Removing rare words harms data diversity