I Can Smell AI from a Mile Away (And the Lecturer
Can Too!)
I’m going to be honest: I am sick of reading and
listening to AI-generated group presentations. Whether it’s an AI image that
looks 'too perfect' or a 2,000-word essay that sounds like a robot, I can smell
it a mile away, and I’m tired of it. This is not to say that I’m not
guilty of using it too. I’ve used these
tools myself, so I know exactly how the 'shortcut' trap works. But if I
can tell it’s fake, believe me, your lecturers can too. Here is how to tell
when someone is being lazy:
- If
your coursework is full of words like delve, tapestry, pivotal, or shaping
the landscape, you’ve been caught. Real students don’t write like
that!
- AI
work often feels "flat" because every sentence is roughly the
same length, whereas real people mix short, punchy points with longer
explanations to give their writing a natural rhythm.
- AI
is always too nice. If you ask it "Who makes the best Rolex in
Kampala?" it will give a boring, balanced answer. Real people have
strong, local opinions; AI doesn't.
- Look
at those AI-generated posters. Are there six fingers on a hand? Do the
people in the background look like melted marshmallows? If the signposts
are written in a fake alien language, it’s a bot.
Use AI to Catch AI
To determine if an image was generated by AI:
- Upload
the image to an AI tool and ask if the image was AI-generated.
- The
AI will search for digital watermarks like a SyncID, or "unnatural
metadata."
- The
AI may also analyse the image for inconsistencies, such as shadows going
in different directions or "mathematically perfect" textures.
If we use AI to do the thinking, we’re the ones losing.
Coming up next: Using Ai without Cheating
