How can I verify Human authorship - a new case study.

Recently I received an email asking me to create a backlink to a lengthy article about online dating scams.  It was a reasonable invite given that we have posted on avoiding online dating scams.  Online dating offers a rich hunting ground for online scammers. 

This is what we did to verify human authorship…

First and foremost I went to the story link provided and read the article.

I checked the ICANN registry of the domain provided in the URL of the Senders email address and learned the web site is registered in the Bahamas, at a legitimate address, to a recognized law firm.

I did a Google search on the company name from the email domain and the first two search result pages to come up were SCAM warning sites, followed by a number of results pages selling IT security services

I searched the name of the person in the Sender’s email address and what popped up?  A single page of a graphic arts designer in Australia.  The profession of the person didn’t really align with the content of the email received.

When using Google to search for the person on LinkedIn (e.g. “Company name Jane Doe LinkedIn) only two pages appeared – the same graphic arts designer in Australia and a mismatched name - nothing else!

I then copied and pasted the email text into AI TruthTeller® and what was the result?  “This is AI”

I copy and pasted the link of the article provided for the backlink.  It went to a legitimate web site, with a picture of the author that wrote the article.  The article was 4,000 words long so I copied and pasted the article into AI TruthTeller.  Result – “This is AI”

I submitted the picture of the “Author” into AI TruthTeller and the result – “This is NOT AI”

The article was “Reviewed By” a separate person.  I pushed the picture of that person through AI TruthTeller and the result – “This is NOT AI”

I copied the “Edited by” picture and pasted that into AI TruthTeller with the following result – “This is NOT AI”

The article itself?  Well it has multiple sections that are duplicated so it looks like the publisher is stuffing content in order to build authority and get better search results.

I do a search of the Author on LinkedIn.  The profile appears real and the Author has 198 connections and last posted on LinkedIn 7 months ago.

The Reviewer also has a profile on LinkedIn with 438 Connections.  The Reviewer “tells stories people like (and Google doesn’t hate.)”

I did a search in ChatGPT of the person who sent me the email: “Who is Jane Doe” at “Company XYZ”?  Nothing.  ChatGPT can find no evidence of a real person with this name associated with the company.

Here’s my conclusion: the email is AI generated and uses the name of an alias or what we nowadays call a fake person.  The company name is real but they have a tendency to show up in SCAM lists.  The author of the article is a real person as is the Reviewer and the Editor.  However, the article itself is AI generated or has extensive AI writing and/or editing.

For comparison, I copied the article title, dropped it into ChatGPT and asked it to write me a 4,000 word article, using the title for subject matter direction and guidance.  I received 4,000 words and more and it had a fair amount of overlap with the article.

Conclusion – the article does not have human authorship and is not a back link I necessarily want to reference.  Having said that, the points in the article are useful even though they appear to be a watered down version of what ChatGPT can produce.  Do I know if ChatGPT is simply regurgitating the article itself, which is already AI generated?  Nope, I don’t but I do know that AI is really good at generating text and the article has a strong probability of being entirely generated by artificial intelligence.

In the end, does it really matter if the text is AI generated or written by a human?  Yes, it does.  The author has lost credibility with me.  This isn’t his research, his thinking or his authorship.  It’s benign if it is purely for entertainment.  However, the topic at hand is dating, online scams and how to recognize the warning signs.  This is about personal safety and the content needs to be validated, accurate and represent lived human experience.  Trusting AI to give dating advice is a little like trying to buy a house with fools gold.  It may look pretty, but in the end, it has little to no value.