# ❓ Frequently Asked Questions [🇬🇧 English](FAQ.md) | [🇨🇳 中文](FAQ_CN.md) --- ## General Questions ### Q: Will this help me avoid AI detectors? **A:** The primary goal is improving writing quality and authenticity. As a side effect, natural writing is less likely to trigger AI detection, but that's not the main purpose. We focus on teaching you *why* text sounds AI-generated and *how* to fix it—not on "gaming" detection systems. --- ### Q: Can I use this for my dissertation/thesis? **A:** Yes, if you're using AI to draft text expressing *your own original research and ideas*. Many scholars now use AI as a writing assistant. The key is that the intellectual contribution must be yours. **Appropriate scenario:** - You conducted research, developed arguments, and have clear ideas - You used AI to help draft text expressing those ideas - You're now improving the naturalness of that draft - ✅ This is ethical use of AI assistance **Inappropriate scenario:** - You asked AI to generate ideas, arguments, or "research" you didn't actually conduct - You're trying to disguise work that isn't your intellectual contribution - ❌ This is academic misconduct --- ### Q: What if I'm not in social sciences? **A:** The core principles apply across fields: - Varying sentence rhythm - Removing mechanical transitions - Adding specificity over abstraction - Natural citation integration Discipline-specific guidance currently focuses on social sciences, but contributions for other fields are welcome! The detection and rewriting strategies work for any academic writing. --- ### Q: Do the scripts require internet connection? **A:** No. All analysis runs locally using Python's standard library. There are: - No API calls - No data sent anywhere - No external dependencies - Complete privacy for your work --- ### Q: Can I use this for languages other than English? **A:** Currently optimized for English academic writing. The principles (sentence variation, specificity, natural flow) may transfer to other languages, but: - Detection patterns are English-specific - Examples are in English - Transition word lists are English-based Contributions for other languages would be valuable additions to the project! --- ### Q: How does this compare to commercial AI detection tools (like GPTZero, Turnitin)? **A:** This tool serves a different purpose: **Commercial detectors:** - Tell you "this is probably AI-written" (yes/no) - Black box scoring - Don't teach you anything **This tool:** - Shows you *why* text sounds AI-generated - Teaches you *how* to fix it - Helps you develop better writing skills - Focuses on learning and improvement, not just detection Think of it as a writing coach rather than a detector. --- ## Technical Questions ### Q: Why Python 3.7+? Can I use an older version? **A:** Python 3.7 is required for features like: - `statistics.stdev()` behavior - String formatting methods - File encoding handling (especially important for Windows) Most systems from 2018+ have Python 3.7 or higher. If you're on an older system, upgrading Python is recommended for security and compatibility. --- ### Q: Can I integrate this into my own tools/workflows? **A:** Yes! MIT License allows: - ✅ Use in commercial projects - ✅ Modification and redistribution - ✅ Integration into other tools - ✅ Use in closed-source projects Just include the copyright notice from the LICENSE file. --- ### Q: The detector says my human-written text is "AI-like." What gives? **A:** The detector identifies *patterns common in AI writing*, which can also appear in: - Template-heavy academic prose - Overly formal writing - Text with repetitive structures If your human-written text triggers warnings, that's actually useful feedback—it means your writing has patterns that make it less engaging. The rewriting strategies will help improve it! --- ### Q: Can I use this with AI models other than ChatGPT? **A:** Absolutely! The patterns apply to text from: - ChatGPT (all versions) - Claude - Gemini - DeepSeek - Any other LLM AI writing patterns are fairly consistent across models because they stem from how transformers generate text (statistical next-token prediction). --- ## Usage Questions ### Q: How long does it take to humanize a paper? **A:** Depends on length and AI detection level: - **Abstract (150-250 words):** 10-15 minutes - **Section (1000 words):** 30-45 minutes - **Full paper (6000-8000 words):** 2-4 hours With practice, you'll get faster as you internalize the patterns. --- ### Q: Should I humanize everything or just problem areas? **A:** Run the detector first, then prioritize: 1. **High concern areas** (red flags) - Address these first 2. **Moderate concern areas** (yellow flags) - Fix if time permits 3. **Low concern areas** (green) - These are fine, focus elsewhere Focus on the most AI-sounding sections for efficient improvement. --- ### Q: Can I use Cursor AI to help with the rewriting itself? **A:** Yes! That's the intended workflow: 1. Run detector scripts to identify issues 2. Ask Cursor: *"Can you help humanize this paragraph using the humanize-academic-writing skill?"* 3. Cursor applies the rewriting strategies automatically 4. You review and refine The skill guides Cursor to apply the principles correctly. --- ### Q: Do I need to cite this tool in my paper? **A:** Not in the paper itself. This is a writing improvement tool (like Grammarly or a writing center). **Do cite if:** - You're writing *about* this tool in a methods paper - You're discussing AI writing tools in your research **Software citation format available in main README** --- ## Ethical Questions ### Q: Is it ethical to use AI for academic writing at all? **A:** The academic community is still developing consensus, but general principles: **Increasingly accepted:** - AI as writing assistant (like a more advanced grammar checker) - Improving clarity of your own ideas - Non-native speakers getting language help **Generally unacceptable:** - AI generating ideas you don't have - AI creating "research" you didn't conduct - Hiding AI use when your institution requires disclosure **Best practice:** Check your institution's AI policy and err on the side of transparency. --- ### Q: Should I disclose AI use to journals/advisors? **A:** Follow your institution/journal policies: - Many journals now require AI disclosure statements - Some advisors want to know; others don't mind as long as ideas are yours - Transparency is generally the safest approach **Sample disclosure:** > "AI tools were used to assist with drafting and improving the clarity of the manuscript. All ideas, arguments, and interpretations represent the author's original intellectual contribution." --- ### Q: What's the difference between "editing" and "misconduct"? **A:** A helpful framework: **Acceptable editing:** - Fixing grammar/clarity - Improving sentence flow - Making your ideas clearer - Getting language help as non-native speaker **Misconduct:** - Having someone/AI write arguments you don't understand - Presenting AI-generated ideas as your own thoughts - Fabricating research you didn't do **The test:** Can you explain and defend every claim in your paper? If yes, you're fine. If no, that's a problem. --- ## Troubleshooting ### Q: The detector script gives an error. What should I do? **Common issues:** **UnicodeEncodeError on Windows:** - Already fixed in latest version - Update your copy from GitHub **"File not found":** - Check file path is correct - Use quotes around paths with spaces: `"my file.txt"` **Import errors:** - Verify Python 3.7+: `python --version` - No external dependencies needed --- ### Q: The Cursor skill isn't activating. Help? **Checklist:** 1. File is named `SKILL.md` (case-sensitive) 2. Located in `~/.cursor/skills/humanize-academic-writing/SKILL.md` 3. You restarted Cursor after installation 4. Try explicitly mentioning it: *"Use the humanize-academic-writing skill"* --- ### Q: Results aren't good enough. What am I doing wrong? **Common issues:** **Not addressing root patterns:** - Don't just remove "Moreover"—vary the whole sentence structure - Don't just add citations—integrate them naturally **Not enough specificity:** - Replace abstractions with concrete concepts - Ground claims in specific literature/examples **Still too uniform:** - Mix short punchy sentences with longer analytical ones - Vary paragraph openings and transitions Review `docs/rewriting-principles.md` for deeper guidance. --- ## Contributing ### Q: How can I contribute? **Ways to help:** - Add discipline-specific examples (economics, geography, communication) - Improve detection algorithms - Translate to other languages - Share your before/after examples - Report bugs or suggest features See [CONTRIBUTING.md](CONTRIBUTING.md) for details. --- ### Q: I found a bug. Where do I report it? **Report on GitHub Issues:** - [github.com/momo2young/humanize-academic-writing/issues](https://github.com/momo2young/humanize-academic-writing/issues) **Include:** - What you were trying to do - What happened vs. what you expected - Error messages (if any) - Your Python version and OS --- ## Still have questions? - **GitHub Issues:** [github.com/momo2young/humanize-academic-writing/issues](https://github.com/momo2young/humanize-academic-writing/issues) - **Email:** pgallerymoon@gmail.com We're here to help! 🎓