AWS Certified AI Practitioner Practice Exam

To prevent a chatbot from returning inappropriate images in response to user queries, what should be implemented?

1. Implement moderation APIs

2. Retrain the model with a general public dataset

3. Perform model validation

4. Automate user feedback integration

The correct answer is: Implement moderation APIs

Implementing moderation APIs is a highly effective strategy for preventing a chatbot from returning inappropriate images. Moderation APIs are designed specifically to analyze incoming and outgoing content for compliance with community standards and safety guidelines. They typically utilize machine learning algorithms trained to recognize harmful, offensive, or inappropriate content. By using these APIs, you can automatically scan the images generated or suggested by the chatbot in real-time before they are displayed to users. This helps in filtering out unsuitable content based on predefined criteria, thus maintaining the integrity and safety of user interactions. Other approaches, while useful in different contexts, do not directly address the immediate need for filtering inappropriate content. For example, retraining a model with a general public dataset may not adequately prepare it for the specific nuances or contexts that lead to inappropriate content generation. Similarly, performing model validation is primarily a process of evaluating the model's effectiveness and reliability but does not actively prevent inappropriate outputs at the point of interaction. Automating user feedback integration can improve future performance, but it relies on post-hoc analysis rather than proactive content moderation.