We’re already in the twilight zone where proof isn’t always real, and real proof isn’t always believed.

Trying to tell the difference between reality and fiction is increasingly difficult, with us steering towards the dystopian existence of – does it even matter?

Well, as almost every Margaret Atwood book (and Orwell’s famed 1984) teaches us – yes, it does. Very much.

A large majority of what we’re seeing online, on our cellphones and in our emails is now either AI-generated, partially AI-generated, or would have been better off if it were AI-generated. But can you tell the difference?

Strangely enough, the more you use AI, the easier it is to spot AI, which brings us to a whole new reality of writing around AI to avoid sounding like a bot.

That’s because, in essence, AI copied us, and now we’re trying to create something completely different so we don’t sound like it.

It’s a weird loop, and one best avoided. But before we descend completely down this meta rabbit hole, here are 6 quick ways to spot if something is written by AI:

1. Grammar is too good

Yip, humans have proofreaders and editors for a reason – we’re not all 100% proficient in grammar. And even when we do check, double-check, and triple-check the copy, it still emerges with some grammatical errors…or merely grammatical signatures that indicate a unique writer’s preference. Not in AI. It’s textbook grammar, it’s lacking in personality. It’s dull.

2. Grating grammar

AI also has its preferences when it comes to grammar and sentence structures. For headlines, the colon (:) is a MUST. Valentine’s Day: 5 Ways to Say ‘I Love You’. Then the introductory sentence is usually accompanied by an em dash – to really land a point. And we can’t forget the ‘not only’… John was not only a handsome critter, he was also kind, amiable, and great at trigonometry. A real package.

3. The rule of three

For whatever reason, AI LOVES listing things in threes – particularly adjectives. It’s used in comedy to make a point, so maybe that’s where the bots learnt it. Either way, when you see text that describes something as well-written, concise and exceptional, you’re likely seeing something that’s AI-generated.

4. Flowery language

AI tends to overuse certain words or phrases that just don’t always flow in everyday language. It’s ADEPT at AMPLIFYING the words LEVERAGED so it can AUGMENT the argument, BROADLY SPEAKING, so this COMMENDABLE, CUTTING-EDGE, STATE-OF-THE-ART technology DEMONSTRATES its proficiency in sounding human-like in this LANDSCAPE (see if you can spot the words).

5. A love of bullets

Look, online text requires interruption through bullet points and subheadings, simply so the modern reader doesn’t lose focus. It allows large quantities of text to be consumed easily. The problem is, AI loves bullet points and subheadings just a bit too much:
· Even
· When
· It’s
· Not
· Necessary

6. Lack of depth

Finally, AI doesn’t try to break down ideas further. It appears to rely on surface-level information and tends to repeat that in a number of different ways throughout the text.

This is particularly (another beloved AI word) evident in thought-leader articles or student essays, where individual thought and critical thinking are required beyond the obvious.

AI is going to get better and better – and it certainly has its place – but hopefully it’s never able to fully replicate the feeling behind a written piece that tells the reader…this nonsense was clearly written by a human.

Don’t have the ZO app? Download it to your Android or Apple device here: