🗣️ My AI usage policy page, on why all of my posts (even on AI) are 100% human-authored, and why I think every website needs an AI usage policy page. (audio now available; 3:39)
This AI Usage Policy page only covered writing, not other uses of generative AI. Here's a new post that summarizes why I don't (yet) use generative AI for creating images, audio, or other artifacts for our newsletters - curious to hear what others think! https://karensmiley.substack.com/p/inconvenient-principles-no-unfair-ai-images
Given all of the buzz around adding audio voiceover to Substack posts, I'm starting to look into using an ethical AI-based tool for cloning my own voice to generate the recordings. Initial article is here: https://sixpeas.substack.com/p/adding-audio-voiceovers-with-ai-ethically Once I complete my evaluation and make a decision, I'll update this page accordingly.
Hi Karen! Question - since you're saying AI use beyond grammar checks, background research, or brainstorming should be disclosed, are you saying those three uses still allow you to say your work is fully human or non-AI generated? I'm not arguing either way on this point - trying to figure out what the standard is and what your thoughts are when I go to make a "real" AI disclosure statement (beyond the no thanks to Pangram statement).
Hi Niki! I’m not here to tell anyone what to say in their AI policy statements. My position in Everyday Ethical AI is that there is value in someone thinking about their values and constraints, writing them down, listing important tasks and activities, and deciding which they want to keep fully human, which to automate (AI or not), and which AI could assist them with. That set of decisions is an AI usage policy, and it’s worth thinking through even if it’s never shared.
As for what to put into a ‘statement’, I think each writer can share what feels important to them for their readers to know.
Some readers have environmental or ethical concerns about using tools based on stolen work. If you use only ethically-developed, environmentally-conscious AI tools (such as the Swiss public AI), that may be worth mentioning.
If it helps, there are guidelines from groups like the Authors Guild and from book distributors like Amazon that say grammar checks etc. do not change whether a work can ethically be labeled human-authored. Those guidelines might help folks decide what to mention in their statements. Or they could just say their work aligns with the Authors Guild guidelines. Or they could use & refer to Credtent’s guidelines on degrees of AI assistance. Lots of possibilities. :)
Let me know if you think expanding this comment into an article with more detailed guidance would be useful.
I definitely think it would be useful! Just to be clear, I'm not looking for anyone to tell me what to do. More trying to gather info because I wasn't sure if there is a common standard or everyone has different thoughts.
It’s a great question, Niki. Probably a lot of writers here aren’t familiar with Authors Guild and other guidelines. I’ll try to write something up. :)
This AI Usage Policy page only covered writing, not other uses of generative AI. Here's a new post that summarizes why I don't (yet) use generative AI for creating images, audio, or other artifacts for our newsletters - curious to hear what others think! https://karensmiley.substack.com/p/inconvenient-principles-no-unfair-ai-images
Given all of the buzz around adding audio voiceover to Substack posts, I'm starting to look into using an ethical AI-based tool for cloning my own voice to generate the recordings. Initial article is here: https://sixpeas.substack.com/p/adding-audio-voiceovers-with-ai-ethically Once I complete my evaluation and make a decision, I'll update this page accordingly.
Hi Karen! Question - since you're saying AI use beyond grammar checks, background research, or brainstorming should be disclosed, are you saying those three uses still allow you to say your work is fully human or non-AI generated? I'm not arguing either way on this point - trying to figure out what the standard is and what your thoughts are when I go to make a "real" AI disclosure statement (beyond the no thanks to Pangram statement).
Hi Niki! I’m not here to tell anyone what to say in their AI policy statements. My position in Everyday Ethical AI is that there is value in someone thinking about their values and constraints, writing them down, listing important tasks and activities, and deciding which they want to keep fully human, which to automate (AI or not), and which AI could assist them with. That set of decisions is an AI usage policy, and it’s worth thinking through even if it’s never shared.
As for what to put into a ‘statement’, I think each writer can share what feels important to them for their readers to know.
Some readers have environmental or ethical concerns about using tools based on stolen work. If you use only ethically-developed, environmentally-conscious AI tools (such as the Swiss public AI), that may be worth mentioning.
If it helps, there are guidelines from groups like the Authors Guild and from book distributors like Amazon that say grammar checks etc. do not change whether a work can ethically be labeled human-authored. Those guidelines might help folks decide what to mention in their statements. Or they could just say their work aligns with the Authors Guild guidelines. Or they could use & refer to Credtent’s guidelines on degrees of AI assistance. Lots of possibilities. :)
Let me know if you think expanding this comment into an article with more detailed guidance would be useful.
I definitely think it would be useful! Just to be clear, I'm not looking for anyone to tell me what to do. More trying to gather info because I wasn't sure if there is a common standard or everyone has different thoughts.
It’s a great question, Niki. Probably a lot of writers here aren’t familiar with Authors Guild and other guidelines. I’ll try to write something up. :)
Thank you!