@KevinB I meant to ask if the user can just post a question to the team without using the Overflow AI feature. Can the user avoid it or it is a forced step? Because that would "match" the scenario above - the user would have to ask to the AI first in order to be "authorized" to post. Which... hopefully is not a thing

@KevinB Do you mean that the Overflow AI feature available on business tier teams works as I described - preventing the user from posting a question directly and forcing them to go thru the LLM asking step?

@Catija therefore the question. The blog as usual is just a chaotic mess of buzzwords, and this was portrayed as the future. I am not accusing anyone here, I just want to clear out any possibility of another "shoe" waiting in the shadow. Or at least be able to look back to this question and point out the "bait and switch" in 6-8 months from now should that nonsense really happen.

@Starship I am willing to think that this is just some stupid miscomunication due to poor images choices and they just refer to some other tools that will provide the ability to post your question to the site if the LLM can't answer it. Your concern about robo review is far worse, especially when the company has shown multiple time that to them AI === LLM and everything else does not exist (including heuristic tools that would be helpful to support manual reviews since a full automated process is illegal under GDPR)

Based on your profile you live in Australia. IF you are lucky enough, your country may not even recognize Stack arbitration clause as legal in the first place - you may want to look into this. Past that, I agree the whole arbitration thing is pretty horrible, especially when you place it in the "Monica" timeframe.

@RyanM oh, sorry, just slang - GAN based generated content. I was just ranting about how these marketing posts the company loves to post label EVERYTHING with the name AI... which most of the time means nothing.

just a little tidbit here, I agree with @RebeccaJ.Stones on a point. It would be probably better if GanGen was called GanGen, LLM were called LLM and the horrid "AI" term was reserved only to the very few tech that actually try to emulate intelligence...

@ShadowWizard That is kinda the issue when AI controls the door. Just ask Hal9000. "I'm sorry Shadow, I'm afraid I can't do that"

You know... I kinda remember some ancient ones that were kinda enjoying the site without really being forced too, despite it still being part of their jobs. I seem to remember some names... Catya, Shog9...... Wonder what happened to them.

the new data dump should have dropped at the start of the new quarter. That would be the start of October (since the August delay was labeled as an outliner due to revised process). Is releasing that part of the plan too?

Congratulation. You have found the Catch 22 trap that lies in basically any LLM ToS right now: there is no assurance that any provided info is correct and you are required to verify any output. So, basically, you can use ChatGPT / LLama / Dragon / whatever to do complex task like summarizing a document to save you the time to read the original .... but you still have to read the original to check that the summary content is correct.

@ColleenV imho, the legal team is fully aware that some of the company actions are not compliant but they are so ingrained in the normal activities that they can't change them .... especially not without loosing the precious $$$ Google probably pays for the data harvesting (noticed how many of the last "community centered actions" seems to be aimed at protecting Google interest? They even have premium user data harvesting access thanks to all the cookie mess). Right now the company is simply considering very improbable that someone manages to get the EU Data Authority on their case