AI has been around for quite some time but it’s with generative AI that it finally found a place for itself in the world’s consciousness. Before that, it was considered underpowered and a cheap alternative.
Generative AI is doing so much better.
But AI could ruin AI.
Have you been noticing how AI-generated content is everywhere? I see articles generated by AI, comments in forums, social posts, and display pics. Everything seems to have an AI flavor to it.
That’s where the ruination is.
You see, AI is excellent because it has been trained on human content. They crawled Reddit, and the Internet, and used stock images and illustrations. Took all your work in every form to create this imitating intelligence.
The trouble is, with the massive influx of cheap AI content there’s less original work to train on.
It’s AI-feeding content to AI, creating a progressively more negative loop where bad AI content trains more bad AI content.
You keep doing that and you have AI that can’t help you at all. It’s just a massive pile of generic crap.
It’s a problem that AI companies will need to confront very fast. How do they keep AI content from making human content inaccessible?
> Journals and magazines are paywalled
> Social media is locked to bots
> No website wants to be crawled by AI
If most of the content on the public Internet is just AI-generated content, there’s not much the next big model can use it for.
Got some answers or observations? I am looking forward to hearing from you.
[link] [comments]