Editorial Standards & Content Methodology

How we create, verify, and maintain content—with full transparency about our human/AI hybrid approach.

Honesty First: How This Content is Created

Let's be completely transparent: Field Guide to AI is created using a human/AI hybrid approach. I (Marcin) provide the context, expertise, direction, and quality control, while frontier AI models (primarily ChatGPT and Claude) assist with drafting, research, and organization.

Here's the honest reality:

I provide human context: Every guide starts with my understanding, experience, and direction about what needs to be covered and how
AI assists with drafting: Frontier models help me write, organize, and explain concepts—they're incredibly fast at producing structured content
I review and refine: All AI-generated content is reviewed, edited, fact-checked, and refined by me before publication
Both can make mistakes: I'm human, AI models have limitations—neither is infallible, which is why we need your help spotting errors

This approach allows me to create comprehensive, high-quality content at a scale that wouldn't be possible with traditional methods alone. But it requires vigilance, verification, and your feedback to maintain accuracy.

Our Content Creation Process

Research & Planning

I identify what topics need coverage based on user needs, AI technology trends, and gaps in existing resources. I research the topic using:

Official documentation (OpenAI, Anthropic, Google, Meta)
Academic research papers and technical blogs
Hands-on testing of tools and techniques
Community feedback and questions

Human Context & Direction

I provide detailed context to AI models about:

What to cover: Key concepts, prerequisite knowledge, common misconceptions
How to explain it: Jargon-free approach, appropriate skill level (beginner/intermediate/advanced)
Real-world applications: Practical examples and use cases to include
Structure: Outline, sections, flow from basics to advanced

AI-Assisted Drafting

Frontier AI models (ChatGPT-4, Claude Sonnet) assist with:

Drafting initial content based on my direction
Organizing information into clear, logical structures
Generating examples and analogies
Suggesting additional sections or considerations

Important: AI output is never published as-is. It's a first draft that requires extensive human review.

Human Review & Fact-Checking

I review every piece of content to verify:

Accuracy: Claims verified against authoritative sources
Currency: Information reflects current AI capabilities
Clarity: Explanations are jargon-free and accessible
Practicality: Examples are concrete and actionable
Completeness: All important aspects covered

Testing & Refinement

Before publication, I:

Test code examples and techniques hands-on
Verify tool recommendations are current
Check that examples actually work
Refine explanations for clarity
Add cross-links to related content

Publication & Ongoing Maintenance

After publication:

Content is marked with publication/last updated dates
Quarterly reviews for accuracy and currency
Immediate updates when AI capabilities change
User-reported corrections reviewed within 48 hours

Our Quality Standards

Every guide and resource published on Field Guide to AI must meet these standards:

✅ Accuracy

All technical information fact-checked against official sources (OpenAI, Anthropic, Google docs). Claims must be verifiable.

✅ Currency

Content reflects current AI capabilities. Updated quarterly or immediately when technology changes.

✅ Clarity

Plain English explanations. Technical terms defined. No jargon without context. Accessible to intended audience.

✅ Practicality

Focus on real-world applications. Concrete examples. Actionable advice users can implement immediately.

✅ Completeness

Substantial coverage (500+ words for guides). Multiple perspectives. Common mistakes addressed. Prerequisites noted.

✅ Independence

No paid placements or sponsored recommendations. Editorial decisions based solely on educational value and accuracy.

Sources We Trust

Content verification relies on authoritative sources:

Official Documentation: OpenAI, Anthropic, Google, Meta AI technical docs and announcements
Academic Research: Peer-reviewed papers from arXiv, conference proceedings (NeurIPS, ICML, ACL)
Technical Blogs: Posts from AI researchers and engineers at leading AI labs
Hands-On Testing: Direct experimentation with tools, models, and techniques
Community Knowledge: Validated insights from AI developer communities and forums

We do NOT rely on: Unverified social media posts, marketing materials, or third-party summaries without checking primary sources.

Acknowledging Our Limitations

Both humans and AI can make mistakes. Despite our best efforts:

I can make errors: Misinterpret documentation, miss updates, or make incorrect assumptions
AI can hallucinate: Generate plausible-sounding but incorrect information
Technology changes fast: Content may become outdated between updates
Complexity is real: AI is nuanced—simplified explanations may sacrifice some technical precision

This is why we need your help. If you spot an error, outdated information, or unclear explanation, please let us know immediately. We're committed to maintaining accuracy through community feedback.

Correction Policy

When errors are reported or discovered:

Our Commitment

48-hour review: All correction requests reviewed within 2 business days
Immediate fixes: Verified errors corrected as soon as possible
Content removal: If content can't be corrected quickly, we'll pull it down until fixed
Update notes: Significant corrections noted in guide with new "last updated" date
Credit: Contributors credited for corrections (if desired)

How to Report Errors

To report an error, outdated information, or unclear explanation, contact me via:

Report Error via builtweb

Please include: Guide URL, specific section with error, what's wrong, and suggested correction (with source if possible)

Content Update Schedule

📅 Quarterly Reviews

Every 3 months, all guides reviewed for:

Accuracy of information
Currency of tool recommendations
Working code examples
Broken links

⚡ Immediate Updates

Content updated within 48 hours when:

Major AI model releases (GPT-5, new Claude versions)
Breaking changes to tools/APIs
Pricing or access changes
User-reported errors verified

Questions About Our Standards?

Have questions about how we create content, verify facts, or handle corrections? Want to suggest improvements to our methodology?