Hi ,

​

One of my greatest fears about generative AI is that, if tech companies don't make it worthwhile for creators to continue producing the content they use to train their models, the most original creators will progressively leave the internet. Then, eventually, all that will remain will be those whose work is produced by generative AI.

Future generations of generative AI will end up being trained on content produced by generative AI, which will then be used to produce new content. And our ideas will never advance beyond wherever they were at that point.

Ultimately, generative AI will become like the Ouroboros - a serpent that devours itself.

[Image: Ouroboros drawing from a late medieval Byzantine Greek alchemical manuscript.]

OK, I'll admit, this nightmare scenario is probably a bit extreme - and I pray it never comes to be. However, it's what sprang to mind while I was recently reading Nate Silver's excellent The Signal and the Noise.

In the introduction to this book, Silver notes that when periods of rapid information growth occur (such as after the invention of the printing press) "it can... take a long time to translate information into useful knowledge, and... if we are not careful, we may take a step back in the meantime."

Generative AI has undoubtedly made it easier for people to access knowledge than ever before and has the potential to transform the world.

But if we're not careful, it also has the potential to devour itself, setting the progress of human thought back in the process.

Talk again soon,

Dr Genevieve Hayes.

All past emails are available in the archive.