CT-AI: Certified Tester AI Testing Version 2.0

By ISTQB · Question Mode

QuestionQ51

Testing AI-Specific Quality Characteristics

A beer company wants to determine how much market recognition its logo has. It plans to do so by monitoring images across various social-media platforms with a pre-trained neural network for logo detection. This particular model was trained by searching for words and by matching colors in social-media images. The company’s logo contains a large word across the middle with a bold blue-and-magenta border.

Which associated risk is most likely when this pre-trained model is used?

Explanation

Pre-trained models can inherit unknown defects, biases, coverage gaps, and other limitations from the data and design used during their original training. Those inherited characteristics can lead to systematically unreliable logo-recognition results in new social-media data.

Community Discussion

No comments yet. Be the first to start the discussion!