Copyright (c) 2026 MindMesh Academy. All rights reserved. This content is proprietary and may not be reproduced or distributed without permission.
2.1.4. Generative AI Workloads
Generative AI creates new content rather than just analyzing existing content. While other AI workloads classify, detect, or extract, generative AI produces original outputs—text, images, code, and more. This is the technology behind ChatGPT, GitHub Copilot, and DALL-E.
Key characteristics:
- Creates original content based on prompts
- Can understand and follow complex instructions
- Powers conversational AI assistants (Copilots)
- Outputs are new, not just retrieved or classified
- Uses large language models (LLMs) and diffusion models
How generative AI differs from other AI:
| AI Type | What It Does | Example |
|---|---|---|
| Classification | Assigns labels to existing content | "This email is spam" |
| Extraction | Pulls info from existing content | "The invoice total is $500" |
| Detection | Finds things in existing content | "There's a face at (100, 200)" |
| Generative | Creates new content | "Here's a poem about spring..." |
Common scenarios:
- Content creation: Marketing copy, articles, product descriptions
- Code generation: Writing code from natural language descriptions
- Chatbots: Conversational assistants that understand context
- Creative writing: Stories, poems, scripts
- Image generation: Creating images from text descriptions
- Summarization: Creating concise summaries of long documents
- Translation: Converting content between languages
- Q&A: Answering questions based on provided context
Generative AI models by output type:
| Model | Output Type | Example |
|---|---|---|
| GPT | Text | Conversations, articles, code |
| DALL-E | Images | Art, product mockups, illustrations |
| Whisper | Text (from audio) | Transcription |
| Embeddings | Vectors | Semantic search, similarity |
Azure services: Azure OpenAI Service provides access to GPT, DALL-E, Whisper, and Embeddings. Azure AI Foundry provides the development environment for building generative AI applications.
⚠️ Exam Tip: Generative AI CREATES new content. If a question asks about producing original text, images, or code—that's generative AI. If it asks about analyzing or classifying existing content, it's NOT generative.
Written byAlvin Varughese
Founder•15 professional certifications