RE: 🤖 AI on Steemit: The DOs, DON'Ts & Community Rules
So, I have lots of thoughts, but many are out of context for a post in the Newcomers' Community, so perhaps I'll try to put them into a separate post in a different context. As an "acceptable use guideline" for newcomers, I'm in general agreement with what you wrote here.
In addition to the uses that you mentioned, I think it's also ok to quote AI output in an article, as long as it's limited, relevant, and acknowledged - just like with any other external source. I also use it sometimes to generate a skeleton outline and/or bullet points and then I'll use biological intelligence to turn the skeleton into full sentences and paragraphs. (like the note cards we all learned to write in grade school)
Bottom line, AI can be an assistant, but not a creator. As you also noted, here:
You are controlling the content, not AI. The point at which AI is controlling the content and not you, you're on dodgy ground.
To me, this is also a key point on moderation from your article:
consistently reaching this level is Downvote territory
Maybe I can't be sure about one post, but if I see 3-5 consecutive posts that are all in the 80-90% range, that's pretty near certainty. If you are suspicious of one post, check the author's posting history.
Another tip for moderators is that if someone leaves comments that are too short for detection, sometimes you can concatenate a bunch of their comments together and evaluate the combined text like a single article.
Coincidentally, I recently saw this, with US copyright guidance pertaining to the use of AI. Really, our use case isn't very different.
key update today from the US Copyright Office for protection of AI-generated works:
- a work can be copyrighted as long as human authorship is a core part of final work - this includes modifying or arranging AI-generated inputs
- humans using AI to generate drafts, edit/refine text, images, music, video, even use AI as a production or creative assistant all qualify for protection
- however just providing text prompts to an AI system is not enough to establish authorship
this aligns with historical precedent where copyright law has supported adoption of new technologies assisting in the creative process such as typewriters, cameras, film recorders, computers, etc
As I wrote this, I considered the same. But "Rules" were requested. So as much as it pained me, I wrote some Rules.
Agreed. As we've discussed before, it's all about transparency. I'll need to update the article to include this.
Good point, I'll include this too. The moderator needs to be intelligent and exclude Diary Games from this (probably).
And this copyright has been discussed elsewhere online - most AI is infringing on the original author's copyright by sharing their content / intellectual property without permission.
I wouldn't be surprised if we see a class-action lawsuit from website owners against AI's unauthorised use of their content.
I agree. It's similar to when Google was sued by publishers and the RIAA. I expect they'll get sued and eventually reach some settlement that involves revenue sharing/royalties.
The new guidelines are about granting new copyrights for artists and writers who use AI, though. It's the output, not the input. Basically, I think Steem's curation decision-makers should follow similar rules to the Copyright office.
This position is hopefully where the article ended up - albeit without knowledge of the Copyright office.
The biggest challenge I’ve found is that whilst I know where my lines of acceptability are, others want a measurable metric. Something which removes subjectivity entirely. Which I keep saying is impossible. But I’ve tried. Always trying. Rarely succeeding.
Agreed. That's why I posted it. I think it reinforces most of what you wrote.
👍
You made a good point. AI should not do the author of the post meanwhile behind you wanted to seek the reference or guide. I always mention this several times in my comments about AI if somebody once you use AI to make research on a particular topic you can also quote the statement an excited the first so that they will know that the text or the idea is being generated by AI. Like you said AI supposed to be a guide but now many people have adopted it to be in the Author of their post what a pity.
Ridiculous post.
What a hypocrisy.
@remlaps
Is he brave enough to also downvote mass AI shitposts pumped out by friends of Justin Sun?
Like this AI-generated spam?
(and many other similar accounts doing the same AI spam)
https://steemit.com/@trafalgar/posts
Or shitpost farming like this?
https://steemit.com/@dodoim/posts
https://steemit.com/hive-150859/@cancerdoctor/416b0b8056e0e
https://steemit.com/hive-196917/@blackeyedm/2025-2-7-14-feat
https://steemit.com/hive-196917/@rtytf2/6v1nb8-ena
and many other of his buddies.
+hundreds of spam accounts owned by them that spam daily and farm the shit out of your beloved Steemit's pool.
Should be consistent with the "rules" and downvote these accounts.