Social Media content creation glitches

  • Thread starter Thread starter DaveC426913
  • Start date Start date
  • Tags Tags
    Social media
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
4 replies · 1K views
DaveC426913
Gold Member
2025 Award
Messages
24,601
Reaction score
8,936
Is see this frequently on Facebook (which is the only social media I frequent).

1702180686719.png


To be clear, what we are seeing here is a Facebook post, but it is a post that is a graphical image - a capture of a block of text (which could have just as easily been just textual content).

Nothing untoward so far...

But there is an additional element here: the graphical image has been corrupted - a few words have been obfuscated.

It happens far too often to be anomalous. It has to be deliberate.

My best theory is that it is an attempt to foil some sort of OCR* content scraper that presumably just grabs content automatically and reposts it as if it's their own.

*optical character recognition

I guess that makes sense, I just didn't know such bots existed. Although, now that I think about it, they would be inevitable.

Any corroborating or alternative ideas?
 
on Phys.org
DaveC426913 said:
My best theory is that it is an attempt to foil some sort of OCR* content scraper that presumably just grabs content automatically and reposts it as if it's their own.
That could be repaired automatically by a spell checker or AI.

The damage appears optical, as if a black/white threshold had been set incorrect during the OCR, or a printer was short of toner or ink.

Why does social media text need to make sense, or be readable without misinterpretation?
Why cut off your nose to spite your face?
 
The text looks like one of the genre of "I'm the only sane man in a world of militant strawman feminists / liberals / racists / jobsworths / other / [delete as applicable]" allegedly true stories that are almost certainly actually "inspired by real events" in the same sense as Hollywood biopics. So I wonder if the text is already stolen from Reddit or somewhere and the errors are weak attempt to prevent OCR-based checks for copyright violation.
 
Ibix said:
The text looks like one of the genre of "I'm the only sane man in a world of militant strawman feminists / liberals / racists / jobsworths / other / [delete as applicable]" allegedly true stories that are almost certainly actually "inspired by real events" in the same sense as Hollywood biopics.
Frankly, I didn't even bother reading the text. Couldn't tell you what the article is about.

Skipping right to the end, you can see it's only a partial story, offered as click bait. I'm so jaded to this kind of clickbait on FB that I just automatically scroll past it - with a flicker of irritation crossing my brow as I do so.

Ibix said:
... I wonder if the text is already stolen from Reddit or somewhere and the errors are weak attempt to prevent OCR-based checks for copyright violation.
That's a new angle. I hadn't thought of that.
 
Baluncore said:
The damage appears optical, as if a black/white threshold had been set incorrect during the OCR, or a printer was short of toner or ink.
I really don't think so.

Again, this happens so often - and so consistent in the effect - that it cannot be an accident.