Skip to Main Content
Today's Hours:

Research Skills Hub

AI: It's...complicated

We're not even referring to students using AI text generators to create text for assignments.

Or how even those NOT using text generators are complaining of their writing being run through plagiarism detectors, and being unfairly labeled as AI-generated.

No, we're talking about whether ANY of us will be able to figure out, going forward, if the text, images, or videos out there on the internet...are AI-generated...or not.

Will we HAVE to use AI tools to detect content made by AI tools? (Is that Irony?)

Asian man with butterfly meme: AI attempting to identify other AI

Well, right now, AI text generation still has recognizable aspects that aren't seen in the same amount in most human writing:

  • Incorrect and outdated information
    • An AI model can't include up to date, more accurate info. that it's never encountered (while being 'trained').
  • Lack of depth and personality
    • AI text bots don't 'understand' what they're generating the way we do when we write - text can be shallow, lacking signs of analysis and depth, and showing none of the personal touches real humans add.
  • Repetitive language
    • Current AI bots are designed to be 'cautious' and 'neutral', and tend to repeat certain words and phrases over and over. (Try getting an AI text generator to make a political statement, I dare you - you'll NEVER get it to sound really partisan.)

From: 

(I think at this point... if you're really being flagged as AI for your own writing... It might be time...to change your writing style.)

 

Tips on checking Images and Videos for signs of AI - Coming Soon!