In PieFed v1.4 it will be possible to label posts as having AI generated content. Labelled posts will get a little badge near their title, similar to how nsfw and nsfl content is labelled. Mods and authors can change this value on a post in the same way they do flair and nsfw, using the little tags button in the bottom right or in ‘More options’ when editing their post.
Account settings for blocking:
Similar to how NSFW works, each user can control how those posts are listed on the home page, etc. The default is ‘Label as AI’, which just adds the badge. People averse to AI-generated content might want to change this to ‘Hide completely’.
Also similar to NSFW, entire communities can be assigned as ‘AI content’ and that will auto-tag every post inside as being AI generated. If the community mods have this value unset (e.g. Lemmy communities, which don’t have this functionality) then the instance admin can manually override the community’s AI Generated setting.
GenAI stands for Generative AI, which I think is descriptive enough
I think their complaint is more that it’s an umbrella term applied a bit loosely (like plain old “AI”). Most people, even non-technical people, seem to instinctively understand that there’s some sort of machine learning (or a novel type of less rigid processing) behind their chatbots and image generators, but most of them also seem to think it’s one artificial entity that does the whole thing. So non-technical ChatGPT users are likely to think all of the different features (like asking it in text to generate an image) are one “consciousness” instead of a text environment and an image environment that sort of interface with each other. And then each of these is made of different parts. It’s all one product (I think we both agree it shouldn’t be one)
I think the person you’re replying to doesn’t like the term for the same reasons I do, it’s quite blatantly just a marketing term, a buzzword to swat away like an angry wasp. But it is the most descriptive and succinct term that most people know. I’m quite conflicted on this. “Taking a few pulls at the slop machines” is a fun way to shit on someone else abdicating their responsibility by using machine generated content en lieu of their brains, mostly in a context where their brains are needed, but “slop machine” is too accusatory. In the same way, I find “GenAI” just a little too permissive or possibly enthusiastic about what is claimed these algorithms even do.
If it was totally down to me I would probably add multiple tags: “Machine generated text”, “Machine generated images”, “Machine generated audio”. It keeps the distinctions I think are necessary (text ≠ thought/analysis, image ≠ artistic expression, etc) and gives us a few more toggles and a bit more statistical information that would be nice.
This comment dragged on a bit longer than I’d set out to write.
I think the issue is that it’s not that much generative nor intelligent
Names the problem without relying on their marketing
That seems vague though: when is something synthetic? If I use something like Context Free to create something, would that count? If you paint or take a picture, but apply some filters / effects? It, to me, feels too general to differentiate between tools that “fill in based on training data, and statistics thereof” and those that do not.
While AI/GenAI is used for marketing, it does draw a clear boundary as to what necessitates the annotation, yet is general enough to include any future changes, while being widely recognized.
No, because what we see is not a randomized output based on statistics of previous content
But 😅 my point is mainly about using a correct technical term, instead of the forced marketing buzzword “AI”