No large language models (LLM) were used in the production of this post.

Inspired by a subtle but clear sign-of-the-times one-line disclaimer at the end of RFC9518’s Acknowledgments (https://www.rfc-editor.org/rfc/rfc9518.html#appendix-A-4)

  “No large language models were used in the production of this document.”
I have added a similar disclaimer to the footer of my homepage:

  “No large language models were used in the production of this site.”
2023 was certainly a year that LLMs took off and stole the hypecycle from #metaverse and #blockchain before that.

Yet unlike those previous two, #LLMs are already having real impacts on the way people create (from emails to art), communicate (LLM chat apps), and work (2023 Writer’s Strike), fueling growing concerns about the authenticity of content, especially content from human authors.

I expect we will see more such disclaimers in the future.

For now, if you blog on your own site with words written by you not #ChatGPT or a similar tool, I encourage you to add a similar disclaimer, and then add your site as an example to the #IndieWeb wiki:
* https://indieweb.org/LLM#IndieWeb_Examples

#largeLanguageModel #LLM #generativeAI #AI

There is the related problem of, when you discover what seems to be an independent site written by a human, how do you know that human actually exists?

For now I’ll mention that XFN rel=met links, published (e.g. metrolls / met-rolls), aggregated, indexed, and queried, can solve that problem. This will be similar to how XFN rel=me links solved #distributed verification on the web (see https://tantek.com/2023/234/t1/threads-supports-indieweb-rel-me and posts it links to).

This is day 48 of #100DaysOfIndieWeb. #100Days

← Day 47: https://tantek.com/2023/365/t1/capture-first-edit-publish-later
→ 🔮

Post glossary:

large language model / LLM

on (ttk.me t5Uh2) using BBEdit