Question

State any two risks associated with Large Language Models (LLMs) that arise from the training process or the training data.

Hint:

Always evaluate LLM outputs critically—training data quality directly affects model reliability.

Answer 1

Step 1: Understand LLM risks.
LLMs are trained on large datasets, which may contain errors, biases, or sensitive information.

Step 2: Bias in training data.
If the training data contains social or cultural biases, the model may produce biased or unfair outputs.

Step 3: Data privacy issues.
Training data may include sensitive or personal information, leading to privacy risks if the model reproduces such data.

Step 4: Misinformation risk.
Models may generate incorrect or misleading information due to imperfect training data.

Step 5: Conclude risks.
Thus, major risks include bias, privacy concerns, and misinformation.