Imagine a world where artificial intelligence shamelessly steals human creativity to churn out endless streams of worthless content – and then picture one clever individual turning the tables in a delightfully rebellious way! This is the fascinating story of a screenwriter who's waging his own quiet war against AI-generated drivel on social media, sparking outrage and laughter in equal measure. But here's where it gets controversial: is this payback justified, or just a recipe for more digital chaos? Stick around to see how he's doing it, and why it might change how you view AI's role in our online spaces.
AI enthusiasts often mimic what real people create, flooding platforms with what critics call 'slop' – that meaningless, low-effort content that looks polished but lacks substance (as explored in this piece on Gen Z's unexpected love for AI-generated art). Yet, not everyone is standing idly by. Scott Collette, a talented Hollywood screenwriter behind the beloved Instagram page 'Forgotten Los Angeles,' spotted an AI-powered Facebook account blatantly copying his historical posts for weeks. Instead of just complaining, he decided to strike back creatively by 'feeding it poison pills,' as he shared in a recent update. These clever edits have led the AI to generate captions so bizarre that the page's followers are melting down in the comments, turning the tables in a way that's both hilarious and thought-provoking.
To understand this better, let's break it down for beginners: AI systems like this one often 'scrape' or copy existing content from the web to train themselves and produce new material. It's like a machine learning from a library of books to write its own stories, but sometimes it gets things hilariously wrong because it doesn't fully grasp context or nuance. Collette noticed the AI was pulling his posts in sequence, so he began tweaking them just before the theft occurred, essentially tricking the algorithm into producing absurd outputs. And this is the part most people miss: it's a simple yet genius tactic that highlights the vulnerabilities in AI technology, showing how human ingenuity can outsmart even advanced systems.
Take this example that went viral: the AI account, known as 'Historical Los Angeles USA,' posted a photo of a devastating flood from nearly a century ago that submerged the city. But the caption? 'A lake made of conservative tears (2025).' The AI then added a lengthy description claiming this 'satirical' take mirrors the 'intense political climate' of the era, where online humor used exaggeration and memes to voice frustrations. It went on to describe the 'lake' as a symbol of 'digital era emotional exhaustion, ideological clashes, and dramatic commentary from the mid-2020s,' portraying it as both a protest and a cathartic release. In reality, this is a classic case of AI hallucinating – generating convincing-sounding nonsense because it lacks true understanding. And the backlash? Pure fury. Commenters blasted it as 'word salad' lacking humor, accused the site of spreading more 'BS captions from this bot,' and some even called it out as 'AI slop' or 'AI feeding off its own excrement.' It's a reminder that while AI can mimic creativity, it often falls flat when human judgment is missing.
Facebook seems to be ground zero for this AI slop epidemic, more so than other major social sites. It's where bizarre creations like 'Shrimp Jesus' – an AI-generated image that went viral – originate, partly because the platform's older user base is more susceptible to photorealistic fakes and authoritative-sounding text. To make matters worse, Facebook's algorithms reward engagement from low-quality AI content, making it financially lucrative for operators to run these accounts. Critics argue this perpetuates misinformation, but supporters might say it fosters creativity in unexpected ways – a point that's sure to spark debate.
Collette's strategy is paying off big time. He explained in an earlier post that by editing his content right before the AI scrapes it, he's ensuring the machine produces laughably inaccurate results. For instance, another post featured a photo that clearly showed a car dealership but was captioned as 'Charlie Chaplin’s Early Los Angeles Home (1905).' Sharp-eyed users quickly pointed out, 'Looks like a 1920s car dealership – he never lived in one.' These floodgates of AI content may be wide open across the internet, but Collette is savoring his small victories. 'It’s been pretty spectacular,' he noted, turning what could have been a frustrating theft into a source of amusement and a cautionary tale.
And this is where things get really intriguing: is Collette's approach a harmless prank, or could it inspire broader resistance against AI overreach? Some might argue it's unethical to manipulate content this way, potentially confusing users further, while others see it as a necessary stand against theft. What do you think – should creators like Collette be praised for their cleverness, or is there a better way to address AI's impact on originality? Share your thoughts in the comments; I'd love to hear agreements, disagreements, and fresh ideas on this evolving topic.
