A few weeks ago, a story broke that unauthorised researchers had been using AI to influence opinions on the Reddit subreddit r/changemyview. Their experiment was simple: could AI-generated comments shift people’s perspectives on certain topics more effectively than posts written by real humans? 

These comments, programmed to be personalised to the each poster’s biographical details, were designed to blend in and persuade. 

The result? Preliminary results showed the AI comments were more effective at changing the minds of Redditors than those posted by humans. The resulting outrage wasn’t just about research ethics: it was about the undetected infiltration of a social media platform whose meticulous human moderation had made users believe it was a bastion of “real human content” away from the bots and slop of platforms like X.

📖 Read more from The Atlantic

And as AI content generation capabilities get slicker, the examples keep coming.

TikTok influencer, Arielle Lorre, recently went public with claims that a company had used AI to generate videos and photos of her promoting products she had never endorsed. It was her face. Her voice. But not her.

📹 Watch the TikTok here

Over in the music industry, it’s been revealed that a North Carolina musician has been raking in millions in royalties from streaming platforms by playing AI-generated songs on repeat to an army of AI bots – a literal heist powered by artificial listeners and artificial creators which took platforms like Spotify and Apple Music 7 years to spot.

🎵 Full story on Substack

It’s unsurprising then, that we’re all starting to second-guess our own abilities to detect “the real” from “the AI”, and that this is spilling over into real life.

Just recently, a Cambridge University student went viral for her eloquent takedown of a right-wing speaker during a public debate. Her performance was praised but for some people it was so good, it was suspicious. She’s had to come out on TikTok to vehemently deny accusations of using ChatGPT to craft and script her arguments. As one comment which has attracted over 450k likes wryly stated: “using ChatGPT is the new witch accusation”.

📺 Watch the clip here

The bigger issue? We are entering a world where we can no longer easily tell:

• If what we’re watching was written by a person or an AI model
• If the voices we hear are real or synthesised
• If the creators we compete with are following the same rulebook
• Whether I really wrote this, or ChatGPT did