🤖 AI Critic & Chat

Get automated quality feedback on your content and ask follow-up questions in plain language.

Overview

The AI Critic is your automated quality assurance assistant. It analyzes your content for common issues across multiple dimensions - factual accuracy, pedagogical quality, bias, and more. When the feedback is too technical, use the AI Chat feature to get plain-language explanations.

🔍

Automated Analysis

AI reviews your content across 6 quality dimensions in seconds.

🚦

Severity Levels

Issues categorized by priority: Must Fix, Should Address, Nice to Have.

💬

AI Chat

Ask questions about any criticism and get plain-language explanations.

🎯

Actionable Feedback

Specific suggestions on how to improve your content.

Using AI Critic

Open the Critic Modal

Click the 🔍 AI Critic button in the content editor toolbar. This opens the critic options panel.

Select Review Type

Choose the type of analysis you need. Each type focuses on different quality aspects.

Wait for Analysis

The AI analyzes your content (usually 5-15 seconds). A loading animation shows progress.

Review Feedback

Read through the structured feedback with severity indicators and specific recommendations.

Ask Questions (Optional)

Click 💬 Ask About This Criticism to chat about any feedback you don't understand.

Critic Types

Choose the right critic type based on your content and concerns:

Type	What It Analyzes	Best For
🎯 Factual Accuracy	Facts, dates, formulas, statistics, references, citations	STEM content, history, technical subjects
📚 Pedagogical Review	Learning objectives alignment, scaffolding, difficulty progression, prerequisite concepts	Any instructional content
⚖️ Bias & Inclusivity	Cultural stereotypes, gender bias, socioeconomic assumptions, inclusive language	Social studies, diverse classrooms, global audiences
📝 Assessment Quality	Question clarity, answer correctness, distractor plausibility, SATA formatting	SATA/assessment content
🔧 Technical Quality	JSON structure, rendering issues, formatting errors, code syntax	Content with code, equations, or complex structures
🎓 Comprehensive Review	All of the above combined	Final review before publishing, high-stakes content

💡 Pro Tip For routine content, use specific critics (Factual, Pedagogical). For important content that will reach many students, always run a Comprehensive Review before publishing.

Severity Levels

Each issue is categorized by its priority:

🔴 Must Fix

Critical issues that should be addressed before publishing:

Factual errors
Wrong correct answers in assessments
Harmful or offensive content
Broken JSON structure
Major bias issues

🟡 Should Address

Important improvements that significantly enhance quality:

Unclear explanations
Weak distractors
Missing scaffolding
Minor bias concerns
Outdated information

🟢 Nice to Have

Optional enhancements for polished content:

Additional examples
Formatting improvements
Style consistency
Enhanced explanations
Accessibility improvements

💬 AI Chat Feature

When AI Critic feedback is too technical, use the chat feature to get plain-language explanations.

How It Works

Click "Ask About This Criticism"

After receiving feedback, click the 💬 Ask About This Criticism button to open the chat modal.

Chat Context is Preserved

The AI already knows your content, the criticism type, and the specific feedback. You don't need to repeat information.

Ask in Natural Language

Type your question as you would ask a colleague. The AI understands context.

Get Actionable Explanations

The AI explains technical terms and provides specific suggestions for how to fix the issue.

Example Questions to Ask

Sample Questions

"What does this mean?"
"How should I fix this?"
"Can you give me an example of what you mean?"
"Why is this considered a problem?"
"What would be a better way to phrase this?"
"Is this really necessary to fix?"
"Can you show me before and after?"

Example Conversation

AI Critic Feedback (Technical):
"The distractors exhibit insufficient semantic plausibility. Cognitive discrimination load is minimal, reducing assessment validity for competency differentiation."

You Ask: "What does that mean? How do I fix it?"

AI Chat Response:
"In plain language: your wrong answers are too obviously wrong. Students can eliminate them without really understanding the topic - they're just picking the answer that 'sounds right.'

To fix this:
• Make wrong answers sound more similar to the right answer
• Use common misconceptions as distractors
• Include partially correct statements

Example: Instead of 'The sky is made of cheese' (obviously wrong), try 'The sky appears blue due to light absorption' (sounds plausible but incorrect - it's actually scattering)."

Quick Question Buttons

The chat modal includes quick buttons for common questions:

What does this mean? - Get a plain-language explanation
How do I fix this? - Get specific action steps
Give me an example - See before/after examples

Recommended Workflow

For best results, follow this quality assurance workflow:

Quality Workflow

1. Generate/Edit Content
         ↓
2. Run AI Critic (start with specific type for your content)
         ↓
3. Review 🔴 Must Fix issues first
         ↓
4. If unclear → Click "Ask About This Criticism"
         ↓
5. Make corrections in the editor
         ↓
6. Re-run AI Critic to verify fixes
         ↓
7. Address 🟡 Should Address issues
         ↓
8. Final Comprehensive Review before publishing
         ↓
9. Preview in student context
         ↓
10. Publish ✅

Tips & Best Practices

✅ Do's

Always run AI Critic before manual review (it catches technical issues quickly)
Use specific critics for targeted feedback
Ask questions when feedback is unclear
Fix 🔴 issues before publishing
Run Comprehensive Review for high-stakes content
Re-run critic after major edits

❌ Don'ts

Don't blindly accept all AI suggestions - use your judgment
Don't skip the critic for "simple" content
Don't ignore 🟡 issues for student-facing content
Don't rely solely on AI - add your expertise
Don't publish without at least a quick critic check

Limitations

While AI Critic is powerful, be aware of its limitations:

Domain-Specific Accuracy: AI may miss subject-specific errors in specialized fields. Always have domain experts review technical content.
Cultural Context: AI may not understand local cultural nuances. Apply your knowledge of your students' backgrounds.
Pedagogical Judgment: AI suggests best practices but doesn't know your specific classroom context. You know your students best.
False Positives: AI may flag issues that aren't actually problems in your specific context. Use judgment.
Language Nuance: For non-English content, AI criticism may be less accurate. Consider having native speakers review.

⚠️ Human Judgment Required AI Critic is a tool to assist your review, not replace it. The final decision on content quality is always yours as the subject matter expert.