To think that mumsnet posters should know that anyone can use AI to analyse their posting history

[LostPollypocket]

I understand why a previous thread on this was deleted and suspect another one will be. I did think the thread was doing something very useful though in making people aware that even anonymous posts may not be so anonymous in a world where anyone (or business, crime syndicate, political party, or government) can use AI to pattern match and analyse everything you’ve ever written on the internet, including places like mumsnet where you might post about very personal topics. If you use a particular word frequently, or semicolons in a particular way, or even respond emotionally to certain types of posts, then even with a name change anyone might be able to build a profile on you and connect it to say, archived facebook data, and know who you are, how you think, and what you posted at 3am in the depths of postpartum depression.

I don’t think this is about censorship? Did you think the same about the Cambridge analytica scandal?

I think that’s an open question, but my money is on it being fairly scalable given enough data (which there is).

Writing kind of is a fingerprint, it can be used forensically.

Damn, Gen Xs can't be smug to Millenials and Gen Z about having no cameras around in our misspent youth anymore because our misspent online adulthood is revealed forever 😂

Before AI came along, I worked out who someone was on here purely from her writing style. So if I can do that, I'm sure AI can.

She didn't put any identifying information in her posts, but I just knew it was her from her choice of words and sentence structure. What is interesting is that she is MUCH MORE extreme in her political views on what she assumes is an anonymous site, than she is in real life.

After several years she did put some personal information in one of her posts, which confirmed my view that it was her. I've never told her that I've identified her, but it's impacted what I think of her.

Surely that's why people name change regularly and tweak details. I don't believe there are as many sets of twins in the world as there are on MN for example

It is. There was a woman on Handbag who was a forensic linguist. She was bloody brilliant, she could spot a namechanger from a couple of sentences. She found me on here once, and PM'd me to ask if I was the same person 🙂

Of course a computer could do it, if trained by experts like her.

Interesting and depressing paper on scalability here - https://aclanthology.org/2026.acl-long.1489.pdf

I am very aware of my own stylistic tics in this thread. It’s exhausting and not much fun to try to suppress them and I don’t think i’ve been successful.

How interesting! What is Handbag?

There are also the people who work out where in the world a photo was taken with terrifying accuracy. So i suppose that’s all up for AI grabs too.

Go take a look at what "they" know about you on Amazon... half of it is rubbish because DH buys on it, I buy on it Dd25 buys on it. But they track buying patterns and surmise personal details about you.

And people already spent money to buy our details when MN was hacked years ago. I still get crap to my old email address that was only used for here. Can see it on the dark web as "associated with old MN username".

Security tip for the day.... DO NOT USE THE SAME EMAIL FOR BOTH USER FORUMS AND BANKING. You'd be surprised how many people do.

for me, the 1st things I think of when i see AI are
Avian influenza
Artificial Insemination

I don’t know what you think my agenda is or why you think i’m the other person? I posted on that thread. My focus there, and here, is somewhat different to the focus of the OP of that thread. AI is an interesting and newsworthy topic and there are issues that affect anyone who uses anonymous internet forums. So I’m interested in discussing it! What else would i possibly be trying to do here…?!

But when it's done forensically, it's looking at a sample of writing known to be written by a particular person, looking at another sample of writing known to be written by a particular person, and determining whether that's the same person. It's a tiny dataset. If one sample of writing were being compared against millions of others, it would pull up too many possible matches for it to be useful.

I hope you’re right but think you’re wrong!

I did find this on scalability. Not millions but tens of thousands, and remember this is all very early work on a rapidly improving technology - https://aclanthology.org/2026.acl-long.1489.pdf

Also all of that is before you add in actual ai agent hacking that can then obtain IP addresses, link usernames, etc etc.

The sentiments you express in that paragraph are exactly those sentiments (fears) expressed by censors wishing to control what information can reach which people. If you don't think your fears about AI are also about censorship then you have a blind spot.

Cambridge Analytica are only visible because they were sued. The scale of data-harvesting is absolutely industrial and we consent to most of it. That's not grounds to assume people are stupid, though.

I don’t think my fears about AI are about censorship but your post hasn’t helped me see my blind spot if that’s what it is. I think just my mention of credulous voters has pressed a particular button for you. I do have opinions about censorship and democracy but I honestly think they’re a tangent here.

Or just not engage with forums like this, other social media and the internet.

We are responsible for our choices.

I do think it's going to get much worse though. If you think about what AI can do now, then try to imagine what AI will be able to do in 10 years, or even one year. Things are moving at an exponential rate (and, yes I mean exponential in the mathematical sense).

On the other hand, there doesn't seem to be much anyone can do about it. I already don't share much personal information online, but I'm sure there's enough. And why should anyone bother? Who knows what bad faith actors of any persuasion would want to do it in the future.

I know I'll be accused of paranoia, but if we could all revisit this thread in a few years I don't think it will have been based on nothing.

But the problem is that what you need to do is not have engaged with forums like this unguardedly ten years ago, or post teenage nonsense on myspace twenty years ago, or on a niche mental health board back before google even existed. All of that data now looks very different in a way that was unimaginable science fiction at the time.

Yanbu I am wary now too x

I love shaming men, but dont want to be denied a passport because of it 😫😫

It's interesting but I still maintain it couldn't be scaled across searching the whole Internet for reliable matches. Combining with IP addresses though like you say could be a different story.
I just think if we did ever reach that stage, people would also be using AI to impersonate and plant writing, so the whole thing would just become so unreliable.

Ok but what about just mumsnet, ovarit and a filtered selection of uk based facebook profiles for a particular age and sex?

And yes to mass red herrings, someone else pointed that out - but our generation(s) are still going to be vulnerable having already generated huge amounts of timestamped analysable data pre AI explosion.

I think you're massively underestimating the speed at which the technology is advancing. Mathematicians are having to deal with the fact that AI has already solved about 300 problems that mathematicians across the globe have being trying to solve for decades, in the space of a couple of years. AI is already producing good quality literature. There's no field, including the ones that people thought would be immune, that won't be massively affected.

Like I said, let's see how things are in a few years.

I think anyone doing that needs to get a new hobby. I mean, what is the actual point in spending the time, effort ( and water) to analyse someone's posting style to try and identify their other use names, or somehow catch them out.
Get a fucking life.

https://www.theguardian.com/technology/2026/oct/17/fake-linda-and-me-mystery-fraudster-stole-my-byline

Somewhat adjacent and ties into how easy it can be with AI tools to commit identity theft / undermine people's credibility etc

And you're right OP, it is too late. Most people's prolific posting histories across social media and forums can easily date back two decades pre AI, and is out there ripe for harvesting. I no longer post on FB if I can help it, here is my only outlet. Now that I'm aware of the dangers and opposed to AI on many levels anyway, and the way of the world has become so brutal, I find very little gain from it all. It ain't fun any more.

I too throw in the odd lie and have different regular typos here than on X or Facebook.

My X style is markedly different, ruder, and my facebook stuff by far more polite.

Well worth developing SM ticks for wach platform you use