aiwg/agentic/code/addons/writing-quality/patterns/common-ai-tells.md

Cognitive architecture for AI-augmented software development with structured memory, ensemble validation, and closed-loop correction. FAIR-aligned artifacts, 84% cost reduction via human-in-the-loop, standards adopted by 100+ organizations.

# Common AI Tells and How to Avoid Them

## The Top 10 AI Giveaways

### 1. The "Comprehensive Solution" Syndrome

**AI Tell**: Everything is comprehensive, robust, innovative, or cutting-edge. **Why it happens**: AI is trained to
sound impressive and complete. **Fix**: Be specific about what the thing actually does.

### 2. The Three-Item List Obsession

**AI Tell**: "The system is fast, reliable, and secure." **Why it happens**: Three items feel complete and balanced.
**Fix**: Use 2, 4, or no list. Or just pick the most important item.

### 3. The Moreover Transition Addiction

**AI Tell**: Moreover, Furthermore, Additionally at the start of paragraphs. **Why it happens**: AI learned formal
writing patterns. **Fix**: Just start the next sentence. No transition needed.

### 4. The Perfect World Problem

**AI Tell**: Everything works flawlessly, all stakeholders are satisfied. **Why it happens**: AI avoids negative
outcomes. **Fix**: Include what broke, what was hard, what you'd change.

### 5. The Hedge Word Haven

**AI Tell**: "can help to", "may provide", "seeks to enable" **Why it happens**: AI avoids absolute statements. **Fix**:
Use direct verbs. It does or it doesn't.

### 6. The Em-Dash Epidemic

**AI Tell**: Multiple em-dashes — like this — throughout the text. **Why it happens**: AI uses them for parenthetical
information. **Fix**: Use commas, parentheses, or separate sentences.

### 7. The Conclusion Compulsion

**AI Tell**: "In conclusion", "To summarize", "All in all" **Why it happens**: AI follows essay structure. **Fix**: Just
stop when you're done.

### 8. The Balance Beam Act

**AI Tell**: "While X has challenges, it also provides opportunities" **Why it happens**: AI tries to present balanced
viewpoints. **Fix**: Pick a side or state facts without the balance.

### 9. The Transformation Fixation

**AI Tell**: Everything is transformative, revolutionary, or game-changing. **Why it happens**: AI amplifies importance.
**Fix**: Most things are incremental improvements.

### 10. The Seamless Integration Illusion

**AI Tell**: Everything seamlessly integrates. **Why it happens**: AI implies perfection. **Fix**: Nothing is seamless.
Everything has friction.

## Structural Tells

### The Wikipedia Opening

**AI Tell**: Starting with a broad definition before getting specific.
> "Digital transformation is the process by which organizations..."

**Fix**: Start with the specific situation.
> "We needed to get invoices online. Paper was killing us."

### The Academic Structure

**AI Tell**: Introduction → Body → Conclusion in every section. **Fix**: Sometimes just make your point and move on.

### The Consistent Paragraph Pattern

**AI Tell**: Every paragraph follows: claim → evidence → conclusion. **Fix**: Mix it up. Some paragraphs are one
sentence. Others ramble.

### The Summary Addiction

**AI Tell**: Summarizing what was just said. **Fix**: Trust the reader. They got it the first time.

## Voice Tells

### The Helpful Assistant Voice

**AI Tell**: "I hope this helps!", "Let me explain...", "Happy to elaborate!" **Why it happens**: AI is trained to be
helpful. **Fix**: Just provide the information.

### The Marketing Voice

**AI Tell**: Everything is an opportunity, solution, or journey. **Why it happens**: AI mimics corporate communication.
**Fix**: Use plain language. Problems aren't opportunities.

### The Academic Voice

**AI Tell**: "It is worth noting that...", "One might argue..." **Why it happens**: AI learned from formal texts.
**Fix**: State facts directly. No throat-clearing.

## Subtle Patterns

### The Perfection Problem

AI never mentions:

- Bugs that took weeks to fix
- Features that users hate
- Technical debt
- Budget overruns
- Missed deadlines
- Arguments about architecture

### The Politeness Pattern

AI always:

- Acknowledges all viewpoints
- Avoids strong opinions
- Hedges statements
- Maintains positive tone
- Provides complete coverage

### The Completeness Compulsion

AI tries to:

- Cover every angle
- List all possibilities
- Provide comprehensive overview
- Include all stakeholders
- Address every concern

## Detection Exercise

### Spot the AI

**Text A:**
> "The authentication system uses JWT tokens with 15-minute expiry. Refresh tokens last 30 days. We store them in
> httpOnly cookies after the XSS incident last year. Still arguing about whether to move to sessions."

**Text B:**
> "Our robust authentication solution leverages industry-standard JWT tokens to provide secure, seamless access control.
> Furthermore, the implementation of refresh tokens ensures optimal balance between security and user experience."

**Answer**: B is AI. Tells: "robust solution", "leverages", "seamless", "Furthermore", "optimal balance"

## The Fix Formula

When your writing sounds too AI:

1. **Add a problem**: What went wrong?
2. **Add specifics**: What technology? What numbers?
3. **Add context**: Why this decision?
4. **Add opinion**: What do you think?
5. **Remove adjectives**: They're usually unnecessary
6. **Remove transitions**: Just start the next sentence

## Real Writing Has

- Typos (occasionally)
- Strong opinions
- Specific numbers
- Tech stack details
- Things that failed
- Inside jokes
- Industry complaints
- Time pressure mentions
- Budget constraints
- Team dynamics

## The Ultimate Test

If you can't imagine a tired developer writing this at 4pm on a Friday, it's too AI.

Real human writing is:

- Sometimes grumpy
- Often imperfect
- Occasionally brilliant
- Always specific
- Never seamless

Remember: You're not trying to sound professional. You're trying to sound human.