AI Slop: Making the Internet Fake-Happy
To anyone with a smartphone and a pulse, the internet is facing an alarming issue. This problem has grown significantly since ChatGPT launched in 2022. Data now backs up anecdotal evidence that AI-generated content is flooding social platforms, including websites.
Key Findings:
- Approximately 35% of all new websites created between 2022 and 2025 are either AI-generated or AI-assisted.
- Online writing is becoming “increasingly sanitized and artificially cheerful.”
- The research team used Pangram Labs’ tools to detect AI, which they found performed well on their tests despite being imperfect.
Symptoms of the Problem:
- AI makes websites more positive.
- AI makes online writing less ideologically diverse.
- The AI-generated content is more consistent and cheerful, with a higher average positive sentiment score.
Conflicting Evidence:
- The analysis found that AI websites were more likely to have similar ideas and viewpoints, making them less ideologically diverse.
- AI was not linked out to external sources or stylistically generic compared to human-written content.
Key Takeaways:
AI is changing the internet in unexpected ways. It’s increasing positive sentiment but reducing diversity and authenticity on websites, potentially making the online experience less diverse and more uniform than before.
- While AI may make content easier to generate for users, it’s important to consider whether this comes at the cost of quality or truthfulness.
- The issue is not just in ChatGPT’s performance but in how AI is integrated into platforms. Efforts should be made to ensure that human oversight remains paramount.
Originally published at Unknown. Curated by AI Maestro.
