AI Slop Is Taking Over Social Media — A Security Expert Reveals How to Spot It

Is your feed even real anymore? In Part 2, I sit back down with security expert and RooVerse founder Ross Thorpe to talk about the AI-generated content quietly taking over social media — AI slop, bots, deepfakes, and how to actually spot it before it fools you. We get into the stats that stopped me in my tracks: how much of what you're scrolling past every day isn't made by a real person, what just a few minutes of chatbot use does to your brain, and the bots making up a huge share of "engagement" on the platforms you use daily. Ross also walks me through RooVerse, the human-only social app he built specifically to keep AI out, and shares the habits he actually uses to stay sane — and real — online. This is Part 2 of a two-part conversation. Haven't seen Part 1 yet? That's where we cover physical and campus safety — go listen to that first. In this episode: 🤖 Why "AI is the average" — and what that means for everything you're scrolling past 🤖 The real percentage of social media images that are now AI-generated 🤖 What 10 minutes of chatbot use does to your critical thinking 🤖 How to spot AI-generated content before it fools you 🤖 Inside RooVerse — the social app built with zero AI content, ever Ross Thorpe is the founder of RooVerse and works with NGS Security, the UK's leading security provider for universities and student accommodation. 🔗 RooVerse: https://rooverse.app/  🔗 NGS Security: https://www.ngssecurity.co.uk/  Sources & studies mentioned in this episode: 🔗 Forbes — AI-generated content now dominates social media images: https://www.forbes.com/sites/bernardmarr/2025/03/10/15-mind-blowing-ai-statistics-everyone-must-know-about-now/  🔗 Cambridge Disinformation Summit: https://www.jbs.cam.ac.uk/events/cambridge-disinformation-summit-2025/  🔗 ScrollAware — the "25 years scrolling" statistic: https://www.scrollaware.com/  🔗 MIT Media Lab — "Your Brain on ChatGPT" cognitive debt study: https://www.media.mit.edu/publications/your-brain-on-chatgpt/  🔗 MIT/Oxford/Carnegie Mellon/UCLA — study on 10 minutes of AI use reducing persistence: https://ai-project-website.github.io/AI-assistance-reduces-persistence/