Original or AI? Check It Out
March 27, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I found “The Psychology of Onboarding: First Impressions Rule the Brain” unusual. I thought one of my team mentioned a similar article at lunch. I do remember her comment to me. She said, “My mom said first impressions count. Now there’s an article explaining how important it is to make a good first impression.” When I read “The Psychology of Onboarding”, it seemed that the article mentioned at lunch was the one appearing in an online publication UXmag.com.
Some people look for the fingerprints of smart software. Thanks, Venice.com. Good enough.
I zipped through the write up and came away with some strong impressions and some hypotheses:
- The write up seemed similar to outputs from Qwen-type large language models. The spacing, the headings, and the assertions without substantive support structures.
- I looked up the LeDoux quotation but, according to the all powerful Google, the quote does not seem to be in LeDoux’ published work.
- Number problems. [a] Not too many ones and [b] rounded numbers
- Weird psychology jargon like “primacy effect”
The big problem for me is the reference to a 2024 Forrester Study. Allegedly this study produced hard data about how perception changes after a first impression. None of the numbers in the write up matches what I could find in a publicly accessible Forrester report. The link does not teleport me to the Forrester report. I ended up on the Forrester.com home page. There’s a similar problem with the Harvard 2023 reference.
Several observations are warranted:
First, I think this write is not particularly credible. I think what my mom told me is, “First impressions count.” I don’t have made up data for my mom’s research to support her statement. As a dinobaby, I know she was correct.
Second, a person looking for information for a high school essay might bumble into this article, recycle it, and find himself or herself demonstrating that short cuts don’t have high reliability.
Third, I asked myself, “Did UXmag.com check this write up before publishing it? Answer: If UXmag.com did, the person doing the review might want to spend a bit more time with the document before posting it.
Net net: Qwen-type LLMs leave fingerprints and hallucinate.
Stephen E Arnold, March 27, 2028
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
