Pamasich

Update 7/31/25 4:10pm PT: Hours after this article was published, OpenAI said it removed the feature from ChatGPT that allowed users to make their public conversations discoverable by search engines. The company says this was a short-lived experiment that ultimately “introduced too many opportunities for folks to accidentally share things they didn’t intend to.”

Interesting, because the checkbox is still there for me. Don’t see things having changed at all, maybe they made the fine print more white? But nothing else.

In general, this reminds me of the incognito drama. Iirc people were unhappy that incognito mode didn’t prevent Google websites from fingerprinting you. Which… the mode never claimed to do, it explicitly told you it didn’t do that.

For chats to be discoverable through search engines, you not only have to explicitly and manually share them, you also have to then opt in to having them appear on search machines via a checkbox.

The main criticism I’ve seen is that the checkbox’s main label only says it makes the chat “discoverable”, while the search engines clarification is in the fine print. But I don’t really understand how that is unclear. Like, even if they made them discoverable through ChatGPT’s website only (so no third party data sharing), Google would still get their hands on them via their crawler. This is just them skipping the middleman, the end result is the same. We’d still hear news about them appearing on Google.

This just seems to me like people clicking a checkbox based on vibes rather than critical thought of what consequences it could have and whether they want them. I don’t see what can really be done against people like that.

I don’t think OpenAI can be blamed for doing the data sharing, as it’s opt-in, nor for the chats ending up on Google at all. If the latter was a valid complaint, it would also be valid to complain to the Lemmy devs about Lemmy posts appearing on Google. And again, I don’t think the label complaint has much weight to it either, because if it’s discoverable, it gets to Google one way or another.

Plus, you explicitly have to opt into this, for each chat you share individually.

I get that it says “discoverable” at first and the search engines are in the fine print, but search engine crawlers get it anyway if it’s discoverable on ChatGPT’s website instead. That term is plenty clear imo.

ChatGPT chats are only public when turned into a shareable chat (which is a manually created snapshot of the chat with a link). And they only show up on search machines if you, after sharing, select the opt-in checkbox for having it show up there.

I don’t know how duck.ai works, but I assume it doesn’t do this.

Also it’s only for AI Windows which requires an AI chip. Considering how anti-AI the fediverse is, I doubt any of the people here would be getting AI Windows in the first place. It’s a non-issue for the users here already from the get-go at the requirements.

You’re delusional if you think Microsoft’s preinstalled crap is at all comparable to what hackers will do with a vulnerable PC.

Also this feature in particular is only for AI PCs which your Win10 PC probably won’t upgrade to.

Except they’re not fighting the fire here, they’re taking away the arsonist’s flamethrowser so he can’t continue making the fire. Without that flamethrower, the arsonist can’t do shit.

Fighting the fire would be petitioning Steam, but the target is the payment processors that pressured Steam on request of Collective Shout.

Symbol is better, as superscript isn’t standard Markdown and isn’t necessarily supported by other software than Lemmy. Mbin for example doesn’t support it.

Not a reason not to use it of course, but it makes the symbol the more preferable choice.

This is incredibly shortsighted.

If you get Collective Shout to stop, another group might pick up where they left off.

The problem needs to be fixed, what you’re suggesting is just making the people currently abusing it stop doing so. That’s not a long term solution.

But americans don’t love Jesus, they shit all over his teachings and would deport him if he came to their doorstep.

Tbf, anything that isn’t AI Windows blocks the feature. Including regular Windows.

People just need to not fall for the scam edition and they don’t have to deal with this shit.

Here’s a question regarding the informed consent part.

The article gives the example of asking whether the recipient wants the AI’s answer shared.

“I had a helpful chat with ChatGPT about this topic some time ago and can share a log with you if you want.”

Do you (I mean generally people reading this thread, not OP specifically) think Lemmy’s spoiler formatting would count as informed consent if properly labeled as containing AI text? I mean, the user has to put in the effort to open the spoiler manually.