💻 Using AI for Content Creation

People are absolutely wetting themselves over the concept of AI content creation with AI-powered tools revolutionizing the way we write. The idea of being able to fill a blog with dozens of articles per day or to have students generate papers with just a few keystrokes is all the rage, and fear. Now you can procrastinate until 5 minutes before the deadline, bash out some ideas, then poof, AI saves the day. Pretty sure that’s not the reality, so I decided to try it out and run a bit of an experiment. I had AI generate a few articles to see what it made.

First, I wrote out a list of topics in an outline. I fed that list into the AI and asked it to use a couple of different styles and tones. The generated content was interesting, and not too bad. It wasn’t amazing, but it wasn’t terrible. It was certainly readable, but it wasn’t something turn in or publish without edits.

In the end, I threw away 80% of what the AI generated. The text it creates is what I would call a “back of the napkin” draft.

• It gets the general idea across
• It was useful to see if I was covering the right topics
• Did I have my topics in the right order?
• Does all my ideas go together?
• What did I leave out?

It definitely helps to identify gaps and provide an example of what an article could look like. It’s like those sample review items or an example resume. These are the kinds of things that make a good resume.

Here’s some quick stats from one of the articles I put together.

Original article length – 543 words

Completed article length – 817 words

There are 179 words that appear in both Text 1 and Text 2

The AI content gave me a ~200 word head start on the article. The pieces I kept were the lines similar to my bullet points.

Getting some AI-generated content served its purpose though. It popped out a completed article so I could see how it flowed from start to finish. I got a complete look at the flow of the topic, which sparked additional ideas.

It’s also an easy way to see different ways of explaining the same topic. In just a couple minutes you can get several variations on a theme. Should this be a serious article or have a light tone? Should it be written for Audience A, or Audience B? As you read over those variations, you can immerse yourself in a subject to determine how you want to write about it.

I also tried getting AI to write a “paper.” I’m not going to lie, it was pretty amazing to input, “What impact has Person A had on Event B” and see how those are linked.

While the information was interesting, and certainly worth more research, the style was a pretty typical “high school 5 paragraph essay” including the phrases, “This paper will demonstrate” and “In conclusion.” Each sample had the same style for its introduction and conclusion. That’s not a bad start as far as content goes, but it’s obviously generated text. Could it have been better if I kept adjusting the tone and style? Probably, but if I spend that much time feeding the AI prompts, I could use that time to edit and polish an article. How would you prefer to spend your time?

It’s incredible to see a computer turn a bullet list into a coherent series of paragraphs. Taking short ideas and turning them into a rough essay or article is a great starting point. Within a few minutes, you can launch straight into editing an article, taking out the pieces you don’t like and filling in with your own tone. It’s much easier to edit and shape a completed article than to build one from scratch. You can immediately overcome that fear of staring at a blank page and not knowing where to start. Fixing a terrible or cliche intro is better than having no intro at all. At least you’re working and the ideas are turning over in your head.

Is AI-generated content the future? I’m sure plenty of people think so. I see it as another tool in the toolbox. It’s no different from using a word processor like Scrivener, or spell-check, or a thesaurus. Having a dictionary app built into the Mac doesn’t make everyone a writer. It’s a helpful assistant. Word processors didn’t kill writing, they enhanced it. The Mac didn’t kill desktop publishing. It made it easier and more accessible. Will AI kill writing? Absolutely not. AI isn’t going to write about current events, or write books that can stand the test of time. AI can mimic Shakespeare or Hemingway, but AI isn’t Shakespeare or Hemingway. Sure, it can be entertaining and knowledgeable, but it still has to get that information from somewhere. People still need to write, edit, polish and judge their writing, not throw in a list and publish an essay without thinking.

Yes, people will absolutely publish AI generated content as their own and fill blogs with articles AI spat out in 30 seconds. People will publish columns, opinion pieces, and their homework using AI. But, these are the same people who used to plagiarize the encyclopedia, so not much has changed.

AI-generated content isn’t the perfect solution to your writing and procrastination needs, but it does have its uses. It can help kickstart the writing process and provide a completed article you can then shape into something great. It can help shape and mold an article into something quite good. It can offer multiple tones to see how they play out. It can write variations on a topic presenting information in different ways. All of that is very useful.

Is it worth using AI to write more articles? I thought it was a fun and engaging exercise. I really enjoyed the process of breaking the article down into bullet points using an outline. Jumping straight to the editing phase really felt like the process was coming together, even with all the edits. It certainly felt faster and like a more fluid process. It’s a very cool starting point for sure, even if it’s a 20% solution. With GPT now getting an endless diet of prompts, that may get bumped up to 30% before you know it. Yeah, I think starting articles with AI might be part of the process moving forward. It certainly doesn’t hurt.