AIBU to think Mumsnet's explanation for that AI prompt raises more questions than it answers?

[JennyJenny72]

I do wonder how many times I've had a conversation online recently without there being an actual person on the other end.

Social media, forums, customer service, online reviews, job applications, even interviews. It seems to be creeping into absolutely everything.

I'm not against AI, but I find it rather depressing that we're getting to the point where you can't be entirely sure there's a human being on the other end.

Is this just going to be normal for the next generation? Will they grow up not really knowing, or even caring, whether they're talking to an actual person?

Or am I just getting old and suspicious?

YABU - Get with the times, grandma.
YANBU - I'd quite like to speak to a human occasionally.

If you're a bot, please start your reply with "Man bot here" so we know where we stand from now on, TYVM.

I think it’s horrible, and really sad. This once felt like a community. It seems completely pointless to be on mumsnet now—the possibility of posters being ai and mumsnet-endorsed drains the life out of the site

I have found that if you swear at bots they sometimes put you straight through to a human. 😳 Though apparently anthropic now want to ban swearing at their precious bots, they would rather it was just people who were abused apparently

Absolutely, talk about diluting away the value. It's like mixing chalk into bread flour or cutting - recreational powders - with ground glass 😠😠

What’s the context for this? Clearly missed something!

Me too!

I am very pro AI in the right context but I absolutely despise the AI generated thread titles on here. Just a race to the bottom.

AI is creeping into our lives whether we like it or not. We have no choice on the larger level; we are mere pawns to play with. We are only at the start of how invasive AI will become.

I am beginning to joke that AI will take over the world - hope it won't but have a horrible feeling that it will be hard to control in the future.

https://www.mumsnet.com/talk/site_stuff/5585556-odd-ai-generated-thread-title-options-in-site-stuff?page=1

A user got a response from mumsnet that was an AI schema that appeared to look like a way of responding to an aibu question

Apparently the AI they use to suggest thread titles went rogue. It also became clear that what you're typing in gets sent to AI before you press submit

It's been picked up by the guardian

This issue has made me think about things, and I suddenly realised how horribly exposed we are by the capacity of AI to read our posts and come to a summary of us as a person - which is likely to be outing. That's already an issue in relation to AI as used by the general public. I don't know whether there are additional perils relating to MN's own use of AI, or its advertising clients' use of MN's AI offers to business. I did read in an article that someone ;linked to on another thread that MN was planning to use AI to imporove its search capacities

And I would like to be able to ask MNHQ what the parameters of this search function would be. Will our change of username be protected? Or will their AI 'know' that all my usernames are me, on the basis of my single registration?

I'd also like to know more about what the effects of deregistering are. Obviously all my posts will remain, but will their connection to my email address and other registration data (and metadata?) be severed? I hope so. In fact I think I will copy these questions into the Site Stuff thread.

In addition to the MN-approved, legal uses of AI, there are the uses that may be made by bad actors, such as aggressive opponents in highly contested topics of online talk. I feel genuinely frightened by the level of exposure we now seem to face.

It's the "please leave a message" senario. You're literally talking to a machine, but because it's actually useful, we do it. Not that it happens much these days, but for 30 years or so it was normal. I think the problem we have now is that most A.I is a solution looking for a problem. It works inside very narrow parameters.

The issue we have now is that the "dead internet theory" isnt just a theory, it's now reality. Facebook reels are now mostly A.I generated crap, being "followed" by bots. This isn't a sustainable business model and will break the link between advertisements and customers. If people believe Mumsnet is a "botfest", then real people will stop using the site. Mumsnet HQ massively fucked up by having anything to do with A.I, it's damaged the brand.

Or perhaps this post has been written by a bot . . . see . . . you just can't be sure anymore.

The USP of Mumsnet has always been that it's a place to have open, genuine discussion with other parents/people with minimal frills and fuss. That's why the layout is so simple and they resisted sparkles and banners and quotes and like buttons etc - the point is the human connection. They are fucking that completely by allowing and embedding AI.

So that’s why there are no Friday night drunk threads. AI’s don’t drink

They used to sound pissed all the time.

Now they've grown up a bit from their student years and are terribly dull and sensible.

There's a thread in site stuff, but it basically turns out that every single thread that was posted was being sent directly to OpenAI in real time to generate the AI title suggestions. Which means that if people were editing their posts to remove any outing details or bits they decided they didn't want to include, it was all getting sent over without the posters knowledge

I read that mumsnet is using bots to start threads or did I misunderstand?

This was about them using bots to respond to threads.

Ah thank you. So to drive engagement. We can’t be sure we’re engaging with humans. Among us indeed.

I have really struggled to keep up with all this. I saw that MN were denying that AI bots were replying to threads but I find that incredibly difficult to believe. In the past couple of months, there have been replies to threads that seemed extremely likely to have been written by AI (as in, generated by, not just a real person using AI to write a response).

As an example, a poster started a thread asking about medical cannabis. Two of the replies on the thread were along the lines of:

‘It seems you are examining your options as to whether to use medical cannabis. Doing some research on this is a great idea and will surely help you come to a good decision’.

It’s just so unlikely that a human would write that as a response because it’s so useless! And there were two replies just like that, from different usernames.

I read a reply on a thread just now and was convinced it was ai before I even knew about this....there's probably far more of them than MN has picked up.

There was also a different thread last week that was so clearly ai assisted if not fully written but the most depressing thing was a different poster, who claimed to be an English teacher, denigrating people pointing this out as being jealous of the post being too well written thus assuming it must be AI. Truly scary that even someone who should have extensive literary experience a) cant tell the difference and b) thinks it is well written.

AI posts are certainly created and posted on MN, but that doesn't mean that MN are doing it themselves. It's a public website, an AI that can access the Internet and interact can post here - MN terms and conditions don't allow it but that doesn't mean that someone somewhere in the world can't set up an autonomous AI agent to sign up as a user and post.

MN will remove and ban the user account if you report that kind of post. There are some areas of the talk section that seem to attract them, which is irritating if those are the topics that you frequent!

@SilenceInside Thank you for explaining that, I didn’t realise that. I really don’t understand AI at all. I’ll make sure to report those sorts of posts when I see them.