First off, one might think I'm very biased writing this article. In fact, I am; I'm a human writer who makes a living off my writing. However, second, I'm also the type who is very open to new technology.

I've been in the business of professional writing since 2005. In fact, with my current platform, I've produced more than 15,000 articles bought by clients, and I was well past 14,000 jobs before that. So, I have a verified knack for writing what clients want.

This evening I decided to test out ChatGPT to see if it even comes up close to a regular writer. As a freelance writer, I'm expected to produce articles on a wide variety of content and topics. Many times it's about information I don't necessarily know firsthand. So, I do a lot of research, find a source with a good command of the topic, and then I use that as a framework template. Smart writers do this all the time as an efficient way to produce content fast.

Of course, we're careful about not copying anything. Writers strive to make sure our work is 100 percent original as it is so easy to trigger a Copyscape flag on anything that looks familiar or similar to content already online. That said, "scraping" is how one gets lots of articles done fast to maintain a good rate of pay. So, in that context, ChatGPT could be very attractive as a way to cut down the "writing" time and boost the earning capacity.

I put it to the test.

Plugging in a topic after signing up with OpenAI, I sat back and watched the bot do its thing. Within 10 seconds flat, it wrote a very well-phrased and grammatically correct paragraph on my question. I was impressed; it was written good. And the task only took a few seconds for anything to respond. It was nothing like the gibberish bot spam from years past. This paragraph easily would have passed muster in a regular generic setting.

So, we move to test 2. Now I want 5 paragraphs on the same topic.

This time, I specified 5 paragraphs in the instructions to the bot, as well as the topic question.

There was a pause, and then the bot started doing its thing. Within 20 seconds, a five paragraph response flowed across the screen. A page worth of material was produced, coherently written and factually correct. My high school teen was with me watching as well, and she was shocked how fast it worked. It was also very enlightening how well the bot put everything together, both in terms of topic sentences for the beginning paragraph as well as supporting sentences. English geeks would be raising eyebrows at the results.

Now, we moved to test 3. Is the content produced original or copied?

Typically, database type robot programs have to use some kind of reference table for their material. As they move through a formula or algorithm, that reference table gives itself away in the output. The question here was whether the writing produce was just a software's fancy version of cut and paste.

Whenever I'm using something suspect, the go-to tool has been Copyscape, but that costs money, and one doesn't need to go that far just to see if a piece of writing is original. Google makes it easier for free. So, I just copied the entire five paragraphs into the search engine and let it rip.

Bingo!

We have phrase matches all over the place. While ChatGPT can definitely put together very coherent paragraphs and even page-lengths papers or more, its strength comes in its vast sourcing from existing material. Many of the artwork AI bots online in Discord now do the same. Phrase after phrase after phrase was an exact copy from other easily-found sources on the Internet. Who is first, ChatGPT or these marketing pages, I don't know, but I can definitely see the material is copied. In this regard, ChapGPT fails. It's not producing a 100 percent original product. That's a game-ender for anyone putting out content for sale against plagiarism checkers or for school assignments.

Be warned. ChatGPT is impressive, but it's not human-grade yet.