How to Use AI to Write Articles in 2023 (What Are the Problems)

These days it is possible to use AI to assist in writing articles or any other type of writing. A typical AI writing tool understands your commands like a human.

For example, you can tell AI to “write an outline for a blog post about cats”. After a few seconds, you get a full-blown outline for your post.

These types of AI assistants offer help in fighting blank page syndrome and streamlining repetitive writing tasks. When used right, you can be more productive, and more efficient, and focus the research or any other groundwork.

The best way to start using AI is by experimenting with different AI writing tools. Also, watch tutorials, read blog posts, and try to use AI wherever possible. This way you might find valuable use cases that might make your job easier and save you time and money.

This comprehensive guide shows you a bunch of ways you can use AI in writing articles and blog posts. I’ve also included the biggest problems with AI writing in this post. Make sure to read those too—there are crucial pitfalls in AI you have to understand.

1. Brainstorm Ideas and Content

Blank page syndrome is common. Sometimes you can’t come up with a topic or next line no matter what.

This is where you can use AI. You can ask AI to come up with topics or subtopics for your articles.

For example, here I’ve used ChatGPT to come up with topics for travel-related topics.

ChatGPT input/output

But make sure to research the topics, though! You don’t want to accidentally choose a too-competitive topic that never ranks on Google.

Make sure to read my tips for writing blog posts that rank high.

2. Write Outlines

Similar to how you can ask AI to brainstorm ideas for your content, you can ask it to write the structure of the article.

For example, let’s pick one of the previously generated blog post topics and ask AI to generate an outline for it.

Writing outlines with ChatGPT

Now, you could take it a step further and ask the AI to complete each talking point in your outline. This way you’d generate a blog post from start to finish only using AI.

But I’ve intentionally left this use case out of the list. There are big problems you need to acknowledge before considering doing that. I’ll talk about those later in this post.

3. Rephrasing

Did you write a paragraph that sounds dull or out of place? You can ask AI to rephrase your text to better deliver your message.

When I write blog posts, I sometimes tend to overthink the sentence or section. This might leave me with a hard-to-read text chapter.

To battle this, it helps a ton to ask AI to write a couple of alternatives to the section.

Take a look at how much ChatGPT AI was able to improve my hard-to-read paragraph:

Rephrasing text with ChatGPT

Of course, if your message is unclear, AI can’t make it much more helpful.

To use AI in rephrasing your content effectively, make sure the text has all the relevant information. The AI can’t read your thoughts!

4. Write Meta Titles and Descriptions

It’s important to make your article’s meta title and description compelling. This way searchers are more likely to open up your articles in the search results.

SEO titles and descriptions with ChatGPT
Example of meta title and description.

Think of a meta title and description like a short commercial for your article. It helps you stand out from other articles on the search engine results.

Sometimes it can be hard to come up with a short yet informative meta title or description. This is where you can use AI to your advantage. Just ask it to write a meta title and description for your article.

For example:

Description example

To use this strategy, ask AI to generate a bunch of samples. Also, don’t blindly stick to the exact words that AI gives you. Make tweaks and modifications as you need.

Also, make sure the meta content is short enough to fit in the search result box. In the above example, the meta description is too long.

5. Auto-Fix Grammar Mistakes

Most of the AI-writing tips you find in this guide are relatively new. This is because the AI writing game has taken big steps in the past couple of years.

But there’s one way you could have used AI in writing for years: Fixing grammatical errors in your text.

You may have heard of a tool called Grammarly. This tool is an example of an AI-powered writing assistant that existed long before the new wave of AI writing software.

Grammar issues fixed by Grammarly

Grammarly is a Chrome extension that uses AI to scan your content as you write. With a tool like this, you can focus less on punctuation and grammar details and pay more attention to writing.

I’ve used the free version of Grammarly throughout my career as a technical writer/blogger.

Problems with AI Writing

Now that you’ve learned how you can use AI in writing, let’s discuss its problems.

One of the major disadvantages people miss is that AI cannot think (yet). This means it cannot do original research.

To write articles that rank high on search engines, you need to provide value to the readers. If you don’t bring anything new to the table, you won’t satisfy your readers. In other words, search engines won’t rank your articles.

I’ve broken down the issues with AI writing into a couple of sections. Let’s take a closer look.

1. AI Can’t Think

The AI writing tools aren’t capable of thinking.

AI is just a mathematical function that takes inputs and returns outputs. Of course, it’s a really complicated function with millions of lines of code, but nothing close to human-level intelligence.

Given your input, the AI algorithm predicts the best output. In a sense, it’s a statistical coincidence if the answer happens to be right.

But because AI is trained with billions and billions of pages of human-written content, it “knows” answers to most of the basic questions. Thanks to this, getting information from AI might be a lot quicker than searching from Google.

2. AI Makes Mistakes

AI makes mistakes. Unfortunately, it does this in a convincing manner that is hard to detect.

There’s no guarantee that the information is right or even close. The less information there is about a topic, the higher the chance of AI making mistakes.

The worse part is that the AI is trained to sound convincing. It can claim things that aren’t true but are really hard to spot unless you’re familiar with the topic.

I tried to write some coding tutorials with AI but failed.

I had to correct a lot of mistakes and false claims made by the AI. It just ended up taking more time than writing from scratch without AI.

Even with nearly 10 years of coding experience, I had a really hard time spotting factual mistakes because they were written so professionally.

Here’s a great example of how AI can mess up even with the basics. I’m trying to ask AI to give a short description for a blog post that fits the search engine result box.

You can read my chat with AI below. (I’ve also summarized the key takeaways from this chat below to save you time).

Issues with ChatGPT

To clarify what’s happening here:

  1. I asked AI to generate a short meta description for a blog post.
  2. The AI gives me a too-long description.
  3. I ask it if it knows the character limit. It clearly does and tells it is 160 characters.
  4. I then ask it to rewrite the too-long description to fit that 160-character limit.
  5. The AI writes too long a description again. This time it also lies about the length.

These types of problems might easily slip through the cracks. The worst part about AI is that it tells lies and does it in a convincing manner.

For example, had I not re-checked the description length, I would have believed it to be 132 characters as the AI told me.

3. AI Lacks Human Experience

This issue is closely related to the “AI can’t think” problem.

When compared to human-level intelligence, AI is nothing but a simple mathematical model.

It has no experience. Instead, it’s taught to predict answers given input.

For example, if you ask AI to write a review of PhotoShop, it surely does a decent job of summarizing what’s been told million times over the internet. But it cannot bring in a unique angle to the review because it hasn’t used the tool—it doesn’t even know what it is.

This is the main problem when writing with AI. You have to do the research and experience part yourself.

Surely, you can use AI to rephrase, summarize, or enrich the experiences you share with it. But you cannot use AI to bring anything new to the table.

4. Plagiarism Concerns

Although AI generates original content, there’s always a chance it doesn’t. If AI misinterprets your intention, it might just start copy-pasting text instead of generating fresh content.

Obviously, this is a big issue—you don’t want to write articles by copying someone else’s work.

Before publishing any AI-written content, check the content for plagiarism. Obviously, there’s no bulletproof way to do that either, but free tools like DupliChecker can give very reliable results.

Wrap Up

AI is coming. The recent developments in natural language processing have made it possible to use AI to write human-like text in seconds.

There are countless opportunities for such AI writing tools. And the best part is they’re only getting better.

You can already streamline some of your writing/content creation workflows with AI.

You can use AI to come up with blog post topics, and outlines, or even rewrite some sections.

Just remember to use AI with care. It lacks human experience and cannot bring anything new to the table. It can only summarize what it has learned from the internet.

I hope you find this guide useful and found some insight into using AI in writing. Happy writing!

Read Also

How to Write a Blog Post That Ranks High