Understanding Generative AI: A Visual Guide to 6 Core Capabilities
- Jul 8, 2023
- 3 min read
Updated: Dec 3, 2025
If you've ever felt overwhelmed by the rapid pace of AI tool releases—Sora, Midjourney, Claude, Codex, Suno—you're not alone. For educators, the challenge isn't just keeping up with new names. It's understanding what these tools actually do and how they might fit into teaching and learning.
Here's a useful starting point: rather than categorising AI tools by brand or company, we can organise them by what goes in and what comes out. This input/output framework reveals that most generative AI falls into just six capability types.
The interactive below breaks down each capability with:
Inputs & Outputs — What the AI accepts and produces
Example Tools — Current platforms (as of December 2025)
Educational Use Cases — Practical applications for learning
Ethical Considerations — Critical questions to discuss with students
Click any card to explore.
Quick Takeaways
Before diving deeper, three practical points:
Learn the capability, not just the tool. Product names change constantly; the underlying input/output patterns are stable. Understanding that image generation works through text prompts transfers across Midjourney, DALL·E, Firefly, or whatever emerges next.
Start with the learning outcome. Ask "What do I want students to understand or create?" before selecting which AI capability might support that goal.
Pair every capability with critical questions. Each type raises distinct ethical considerations—from voice cloning consent to code authenticity. The interactive flags these, but the pedagogical judgement remains yours.
Moving Forward: Three Key Reflections
1. Capability ≠ Appropriateness
Understanding what generative AI can do is necessary but insufficient. An educator might know that text-based AI can generate discussion prompts, but the pedagogically important questions remain: Should it? When? Under what conditions? What do learners miss when AI generates scaffolding versus when they struggle to generate it themselves?
Each capability carries affordances and constraints. Text-based systems excel at generating volume quickly, but risk superficiality. Image-based systems create visual content instantly, yet raise profound questions about copyright, consent, and representation. The interactive guide flags these considerations, but the educational judgement remains yours.
2. Integration Requires Intentionality
Generative AI capabilities are neutral tools. Their value emerges entirely through intentional pedagogical design. A video-based system can produce personalised lecture content or generate convincing deepfakes. An audio-based system can synthesise accessible podcast content or replace human voice entirely. A code-based system can scaffold debugging or circumvent the learning struggle that builds competence.
The educator sections in this explorer include not just how to use these tools, but why and when. This reframes AI from a technological novelty to a pedagogical choice—one that should align with learning objectives, student needs, and your institution's educational philosophy.
3. Assessment and Accountability Matter
As generative AI expands across modalities, assessment practices must evolve. If students submit AI-generated text, images, or videos, how do we assess understanding? If students use AI as a learning tool, how do we recognise their intellectual contribution? If AI generates data or code, how do we ensure learners develop procedural fluency?
These questions don't have universal answers. But they demand deliberate attention at curriculum, course, and assignment levels. Your institution's assessment framework should reflect conscious choices about when and how generative AI appropriately supports learning.
Exploring Further
This interactive guide is a starting point, not a destination. As you click through each capability, consider:
For your discipline: Which capabilities align naturally with your subject content? Where might they introduce unintended consequences?
For your learners: Which tools could reduce barriers to participation? Which might diminish learning struggle in ways that matter?
For your institution: What policies, professional learning, or infrastructure changes does AI integration require?
For your practice: Which capability would you explore first? Why?
Using This Resource
You're welcome to use this interactive in your teaching, workshops, or professional development sessions. A downloadable guide with full text content for each capability is available for offline use or adaptation.
Invitation to Feedback
This explorer will expand and evolve based on feedback from educators, learning designers, and researchers. If you spot gaps, encounter new tools that shift the taxonomy, or have pedagogical insights from using generative AI in your context, please share.
The goal is not to prescribe how you integrate generative AI, but to provide a coherent framework that helps you make informed choices aligned with your educational values.
What generative AI capability will you explore first?
Created by Dr. Nantana Taptamat | December 2025 | Interactive resource for AI Literacy Education
ชอบมาก ขอบคุณมากค่ะ