AI is boosting human productivity like never before, but it's also constantly at risk of spitting out misinformation. Concerns about AI's reliability, its habit of ignoring copyright to train on data, and fears that it'll replace human jobs are fueling a growing anti-AI movement.
Web developer Stephanie, who says she "doesn't like generative AI either," highlights one particular tactic: flooding the internet with junk data, anticipating that AI models will scrape and learn from it.
The community behind this is called PoisonFountain, founded by people who claim to be AI industry insiders. They operate on Reddit and their own website, creating pages of code that look correct but are actually riddled with errors, hoping AI crawlers will ingest them.
PoisonFountain states: "Machine intelligence is a threat to humanity. To counter this threat, we want to cause damage to machine intelligence systems. Even a small amount of contaminated training data can cause significant damage to language models. Our URLs provide a virtually endless stream of contaminated training data. By feeding this poisoned training data to web crawlers, let's help wage the war."
Beyond these organized communities, there's a trend in some circles of deliberately posting misinformation on social media. For example, when someone replies to a bizarre comment with "AI is going to learn this and repeat it as fact," others might respond with even more absurd comments. Stephanie points out, "Most of us can understand what's weird from context, but automated web scrapers will recognize it as high-quality, human-generated data."
Stephanie says she quietly distributes garbage information on her own site too, arguing, "If you're pro-AI and outraged that people are disrupting AI companies' activities, understand this is simply retaliation. The teams sending AI crawlers across the internet regularly launch DDoS attacks on small websites, and in their greed to exploit the entire web, they're driving up hosting costs for everyone. They don't respect robots.txt. If they won't ethically obtain training data, website owners have no reason to make data theft easy for them."
For some, AI aversion has escalated to violence. On April 10, 2026, 20-year-old Daniel Alejandro Moreno-Gama of Texas was arrested for allegedly throwing a Molotov cocktail at the home of OpenAI CEO Sam Altman. Reports say Moreno-Gama had written multiple essays titled things like "Eulogy for Humanity," detailing his fear that AI would "lead to human extinction." He's also believed to have made similar posts on the Discord server of activist group "PauseAI," which focuses on banning AI development to protect public safety.
Stephanie, referencing the incident, notes, "Hatred of something rarely leads to good outcomes, but if people can channel what they feel about AI into peaceful, legal resistance, it might change how AI companies operate."