A company wants an ML model to analyze social media reviews and label each as neutral, positive, or negative. Which modeling approach is appropriate for this requirement?

Classification is the appropriate modeling approach because the goal is to assign predefined categories (neutral, positive, negative) to input data (social media reviews). This is a classic example of a multi-class text classification problem. Open-ended generation involves creating new text, which is not the objective here. Text summarization condenses text into a shorter version, which is also not the goal. Machine translation converts text from one language to another, which is irrelevant to labeling sentiment.