Editors and AI, Part II: AI in Editorial Software—Which Editing Tools Use AI and Which Don’t

In part I of this series, I looked at what artificial intelligence (AI) actually means for editors. I explored how AI is an umbrella term—like using “vehicle” to describe everything from skateboards to cargo ships—and how understanding these distinctions helps us have clearer conversations about AI’s role in editing. We also learned about the two types of AI that editors will encounter: machine learning and generative AI.

Today, let’s look at the editorial software that professional editors use every day. Which tools actually use AI? Which don’t? And why does it matter? Understanding these differences helps us choose the right tools for each editing task and explain our processes to clients.

Traditional Editorial Tools: The Rule-Based Foundation

Traditional spell-checkers, grammar checkers, and PerfectIt don’t use AI—instead, they use predefined rules to identify errors. Think of these rules as sophisticated “if-then” statements: “If a word isn’t in the dictionary, then flag it as misspelled.” Similar to Word macros, these tools are reliable for editors precisely because they follow rigid rules.

Traditional Microsoft Word Features

• Basic spelling and grammar checkers use traditional rule-based checking
• AutoCorrect follows preset patterns
• Find and Replace uses rule-based pattern matching

These features work the same way every time because they follow predefined rules. When you type “teh,” AutoCorrect changes it to “the” because it’s following a specific instruction, not making an AI-powered decision.

PerfectIt

• Uses rule-based technology for consistency checking
• Relies on customizable style sheets
• Works through pattern matching to identify inconsistencies based on set rules

PerfectIt’s developers have been clear about why they don’t use AI: Rule-based systems are more reliable, especially for copyediting. When you need to ensure every instance of the word “healthcare” is spelled as one word to follow your style guide, you want a tool that follows exact rules, not one that makes probabilistic suggestions.

Editorial Tools That Use Some AI Features

Unlike rule-based systems that rely on their own internal “if-then” statements (e.g., “if a lowercase letter appears after a period, flag it as an error”), machine learning models are trained on vast datasets of text. This allows them to adapt to how language evolves, rather than being limited to their pre-programmed rules. However, their output still depends on the patterns seen in their training data, which means they don’t “understand” language like humans do but instead predict the most likely suggestions.

Many editors find that machine-learning tools are less reliable than traditional rule-based systems, and these tools often give strange or “just plain wrong” suggestions. This is because of how these tools are trained—it all depends on their dataset, which changes and grows over time. Like with any editorial software, professional editors will need to check these tools’ outputs closely.

Many of our familiar tools now incorporate AI features alongside their traditional functions. Think of these as hybrid tools—they combine rule-based reliability with AI-powered insights.

Microsoft Word 365’s Editor Feature

• The Editor pane (the blue quill icon) uses machine learning to make advanced writing suggestions
• Analyzes the document’s overall context and style
• Suggests ways to make sentences clearer or shorter
• Includes a “refinements” section that suggests alternate wordings

Grammarly

• Combines traditional rule-based checking with machine learning and generative AI
• Different features may use different technologies
• Can adapt to someone's writing style over time

Adobe Acrobat

• Core PDF creation and editing features are rule-based
• Uses machine learning for specific features:
  • Recognizing and converting scanned text
  • Organizing documents and detecting form fields
  • Improving search functionality

Remember: Whether a tool uses AI or not doesn’t determine its usefulness. Traditional Word features like Find and Replace remain invaluable for editors precisely because they follow predictable rules. Meanwhile, AI-powered tools like Editor can help identify broader issues but require more editorial judgment to evaluate.

Generative AI Tools

Generative AI is what most people think of as “AI” at the moment—tools like ChatGPT and Claude that generate new content by analyzing patterns in massive datasets of existing text.

Unlike rule-based systems, which follow predefined instructions, or traditional machine learning tools, which identify and suggest patterns based on existing text, generative AI can produce entirely new content that mimics human writing. This is achieved through large language models (LLMs) that predict the most likely sequence of words based on the input they receive.

While generative AI is sometimes surprisingly good at creating content that sounds natural, it still relies on patterns in its training data and lacks true understanding or original thought.

Editors who use generative AI typically experiment with ChatGPT, Claude, and Perplexity, then choose the one or two they like the best. I’ll cover these tools in more detail in future posts, but keep in mind that their capabilities are constantly evolving.

Some developers are creating tools that rely on generative AI (usually ChatGPT) to help editors with copyediting and line editing. I’ve tested a few of these, but so far, only one functions decently:

editGPT

• Uses generative AI (ChatGPT) to provide suggestions for wording, spelling, grammar, conciseness, and tone
• Can work within ChatGPT (as a browser plugin) or as a web application
• Exports changes as .doc or .docx (web version only)

While tools like editGPT can be powerful, they require careful oversight. Unlike rule-based tools that follow consistent patterns, AI-first tools can be unpredictable and need thorough checking of their suggestions.

Also, keep in mind that you should never input a client’s work into a generative AI tool without a) their written consent and b) following the highest standards for data privacy and protection. I’ll discuss some of the ethics of using AI in a later post.

What This Means for Your Editing Work

When we debate AI’s impact on editing, we need to be specific about which technologies we mean. For example:

• A rule-based grammar checker (not AI) is fundamentally different from a generative AI system that generates new text
• The machine learning in modern editing tools (a form of AI) presents different challenges than generative AI

Understanding these distinctions helps us make informed decisions about which tools to use and how to use them.

The key takeaway? When we say, “I don’t use AI in my work,” we’re not being clear enough. Many editors already use tools powered at least in part by AI machine learning, like the Editor pane in Microsoft Word. And if you started using Grammarly before 2022, you may have unknowingly used its generative AI features since then. (Wouldn't it be nice if these tools gave us a heads-up as to which features use AI?)

Similarly, asking clients to tick a box certifying that they have not used AI with their work doesn’t actually mean that much, and leads to the all-or-nothing thinking that causes so much shame and fear around this subject. 

As you’re learning in this series, AI can mean many things. The real questions for this new era are: Which specific technologies are we and our clients comfortable using? For what purposes? And with what limitations? The more we’re able to have open conversations and get into the specifics with colleagues and clients, the more prepared we’ll be for whatever comes next.

In my next post, I’ll explore how generative AI tools work, why they make mistakes, and what this means for your editing practice.

Recommended Reading

“Why We’re Not Adding AI to PerfectIt.“ Intelligent Editing Blog. Accessed January 13 2025.

“Microsoft Editor checks grammar and more in documents, mail, and the web.” Microsoft. Accessed January 20 2025.

“Transforming How the World Communicates Through AI.“ Grammarly. Accessed January 18 2025.

Previous Posts in My “Editors and AI” Series

• Editors and AI, Part I: What Is AI? A Primer for Editorial Professionals

This post was published on January 27, 2025.