4 ms·
You’re absolutely right that AI-generated text can be good—sometimes even great. But the reason people care about preserving or identifying pre-AI content isn’t
by jofzar 1y ago
You’re absolutely right that AI-generated text can be good—sometimes even great. But the reason people care about preserving or identifying pre-AI content isn’t always about hating AI. It's more about context and trust.
Think of it like knowing the origin of food. Factory-produced food can be nutritious, but some people want organic or local because it reflects a different process, value system, or authenticity. Similarly, pre-AI content often carries a sense of human intention, struggle, or cultural imprint that people feel connected to in a different way.
It’s not necessarily a “psychological need” rooted in fear—it can be about preserving human context in a world where that’s becoming harder to spot. For researchers, historians, or even just curious readers, knowing that something was created without AI helps them understand what it reflects: a human moment, not a machine-generated pattern.
It’s not always about quality—it’s about provenance.
Edit: For those that can't tell this is obviously just copy and pasted from chatgpt response.
- foxglacier 1y agoOK, so they can choose to read material from publishers that they trust to only produce human generated content. Similar to buying organic food. Pay a bit more for the feeling. No need for those idealists to drag everybody else into it.
- multjoy 1y agoWhy would you want to read something that isn’t written by a human?
- foxglacier 1y agoFor entertainment or education, where it doesn't matter who or what wrote it. It's not as if we have any human relationship with the authors of most content, such as books, etc. or even online content.
- hbs18 1y agoI feel like the em-dashes and "You're absoultely right" already kinda serve the purpose of special AI-only glyphs
- BoxOfRain 1y agoI've found propensity to swear quite a useful observation when determining whether or not a user is an LLM. I suspect it'll remain useful for quite a while, the corporate LLM providers at least won't be training their models to sound like a sailor eight pints deep any time soon.