Given that AI models are trained on Reddit data, do you think someone has already been ‘shittymorph’-ed by now?

Exploring AI Knowledge Through Reddit-Inspired Communication Styles

As artificial intelligence models are increasingly trained on vast datasets—including content from platforms like Reddit—one intriguing question arises: Has anyone, somewhere along the line, experienced a “Shittymorph” transformation within these AI systems?

This curiosity prompted me to experiment with AI responses, specifically by asking the Gemini chatbot to reply in the distinctive “Shittymorph” style—a colloquial, often humorous, and intentionally unpolished manner of communication prevalent on certain Reddit communities.

The results were not disappointing; Gemini adeptly captured the tone, demonstrating how deeply these models can adopt niche internet lexicons when prompted.

This exploration opens up further questions about the limits of AI understanding. Could delving into more obscure or specialized Reddit subcultures reveal the breadth of a model’s knowledge? By examining how well AI can emulate or comprehend less mainstream communities, we might better gauge the depth of their training data and contextual awareness.

If you have ideas or insights into how AI interacts with niche internet lore, share your thoughts—we’re just beginning to uncover the extent of what these models can learn and replicate.