AI-900 Question 250
Select 3A company wants to create a generative AI solution that can automatically generate product descriptions for its e-commerce platform. Which features of generative AI would be most relevant to this solution?
- A
Ability to generate natural language text based on input prompts
- B
Capability to create custom machine learning models without any training data
- C
Support for multimodal inputs like combining text and images
- D
Integration with pre-trained language models for faster development
- E
Ability to ensure generated content is always factually accurate
Show answer and explanation
Correct answers: A, C, D
Explanation
A generative AI solution for generating product descriptions would need to leverage natural language generation capabilities, support for multimodal inputs (e.g., text and images), and pre-trained models to reduce development time. However, it cannot guarantee factual accuracy and does not inherently work without any training data.
- A. Correct.
Generative AI excels at generating natural language text, which is a key requirement for creating product descriptions.
- B. Incorrect.
Generative AI typically requires training data or pre-trained models for effective functioning. It does not inherently build models without any data.
- C. Correct.
The ability to handle multimodal inputs, such as combining text and image data, is useful for scenarios like e-commerce where images of products might be used alongside text.
- D. Correct.
Pre-trained language models, such as OpenAI's GPT, allow for faster and more efficient development of generative AI solutions.
- E. Incorrect.
Generative AI cannot guarantee factual accuracy in all cases, as it generates content based on patterns in the data it has been trained on, which may not always align with facts.