I Hate This Stupid Baby Platypus

Why a fake baby platypus should worry you more than you think.

One of my favourite rabbit holes is the world of mis- and disinformation online—especially on social media. The sheer impact that disinformation has already had on consumer trust, elections, public health, and society at large genuinely astounds me.

Take, for example, the photo of the ‘baby platypus’ that circulated widely a while back. Thousands of reacts, shares, and heartfelt comments poured in, all celebrating this adorable creature. Except, of course, it wasn’t real. It was a sculpture—a beautifully crafted wool and clay piece by Russian artist Yuliya Leonovich, the photo stolen from her art accounts.

Seems harmless, right?

Except it isn’t.

The Data Trail You Don’t See

Those innocent shares and comments aren’t just signals of awe or wonder. They’re often data points, breadcrumbs for content farms and disinformation networks to follow. Every like and share helps these operations map out who is most likely to believe something without fact-checking—who is most emotionally reactive, who is less digitally literate, and who might be prime targets for more sinister campaigns down the road.

It’s the digital version of boiling a frog. One minute you’re sharing a cute baby animal, the next you’re being served a politically charged deepfake video designed to manipulate your worldview.

And here’s the kicker: people already struggle to spot AI-generated imagery from real photos.

This Is Not Theoretical

Meta itself has taken down multiple disinformation networks that used AI-generated profile pictures to push political propaganda, conspiracy theories, and voter suppression content. From pro-China operations creating fake American personas, to political astroturfing in Argentina, AI-generated faces and fake accounts are actively shaping discourse on Facebook and Instagram right now.

And now, with OpenAI’s Sora landing in the UK—giving anyone the power to generate hyper-realistic AI video—the line between real and fake just got even blurrier.

A quick scroll through Sora’s showcase already reveals that users are creating exactly the kind of content that low-digital-literacy audiences are most likely to fall for: a baby in a tiger onesie being licked by a tiger, a tiny deer cradled in a hand, footage that could easily pass for a war zone—except none of it ever happened.

Why Visuals Hit Different

This matters because visuals bypass the critical thinking we might apply to text. They activate our emotional responses first. Studies show that vulnerable audiences, particularly those with lower education or limited experience navigating online misinformation, are far more likely to believe and share false visuals—especially when those visuals confirm their existing beliefs.

And this is where AI and divisive political content intersect.

Once you know someone is quick to trust a baby platypus, it’s not hard to target them with a doctored image of a ballot box being stuffed, or a deepfake video of a politician saying something inflammatory. It’s confirmation bias on steroids, and it’s already happening—especially in private spaces like WhatsApp, where fact-checkers can’t easily intervene.

What Needs to Happen

AI-generated content—whether it’s a whimsical platypus or a fake piece of wartime footage—must be clearly and visibly marked as synthetic.

Platforms like Meta are moving in this direction with watermarking and labeling policies, but enforcement will need to be fast, consistent, and multilingual to have any meaningful impact. And it’s not just on the platforms—we need a cultural shift toward digital literacy, teaching people how to spot synthetic content and question what they see online, especially when it seems too good—or too outrageous—to be true.

Because the next baby platypus might not be so cute.

Further Reading