Thread 3: A corpus-assisted discourse analysis of linguistic transphobia on Mumsnet

That unfortunately makes it sound as if it's all about paying Mumsnet and to hell with the consequences if people get doxxed in the process.

Cold hard cash would be good.

Or at least making all the premium features free ? Nah… cold hard cash.

I'd never noticed it was the MN logo on the mug. I absently thought it was a blue flower 🤯

Here's your payment. You even get a Mumsnet mug!

I'd like to think a percentage of any payment would be used for the good of MN users.

Perhaps MN might fund information gathering and focus groups from its members in order to be a stakeholder to relevant consultations about maternal and neonatal health, women's health, mental health etc.

'There is nothing wrong with mining content for data, but it has to be properly regulated and creators must be compensated'

Ooh, are we going to get paid?

https://www.theguardian.com/commentisfree/2024/sep/28/mumsnet-ai-google-openai-publishing-copyright

No mention of Aston in here.

Am not surprised.

But realise that in some foolish, unlike me optimistic way, I had hoped they would review how they work. They could have conceded there were issues, and as a result of MNHQ contacting them, they have decided they will in future be much clearer about how they collect data, and make it prominent of their website, not as a footnote on a document.

They seem to be taking a bullish, unrepentant tone.

That is singularly underwhelming.

There's an update from Justine on the Corpus 2 thread.
https://www.mumsnet.com/talk/site_stuff/5072884-corpus-2?reply=135821583

Deleted for breaking talk guidelines I mean. Suspect MNHQ would have been erring on the side of caution.

Thread 1 there are plenty of deleted posts.

I've looked back through the threads and can't see any deleted posts. I thought that,if you asked for a post to be deleted, MN leaves a notice saying that they've deleted it at the request of the user. Am I looking in the wrong place or have MN deleted evidence of a post existing at all rendering it completely invisible? Can someone post a link to a relevant page? Thanks.

looking back at that discussion on 20 April, there are quite a few deleted posts now!

??

Have posters asked for posts to be removed, or is MNHQ "cleaning house" in relation to challengins Aston.

It would be really good to create a summary of what happened with relevant screenshots etc, and I keep meaning to do this but don't have any time.

The sort of job that might suit a PhD student with some unexpected time on their hands, for example. Anyone know someone like this?

Not sure if this has been shared already, but this is an article about FoLD

Yes, it was discussed thread 1 on the Friday night before the event was shut down, but just incidentally, looking back at that discussion on 20 April, there are quite a few deleted posts now!

I mean, FFS.
What kind of logic goes - I know, we want to look at despicable crime against children, we will use a corpus of discussion about infertility and adoption to test our model?!?

And then - okay, this might have not been the best idea, but we still want to hold onto the FWR threads (where women have expressed concerns which have been upheld by the Cass review) and sit them alongside the despicable crimes?

Not sure if this has been shared already, but this is an article about FoLD, the repository where the MN dataset was stored. https://publications.aston.ac.uk/id/eprint/43719/3/Aston_Forensic_Linguistic_Databank_FOLD_001_petykoetal.pdf

Worth remembering that the MN corpus was a controlled dataset, if I remember correctly.

Did the people filling in Arabic and Kurdish questionnaires know where their data was heading? What was their crime?

The company we keep, eh? 😜

I've reported your posts, @GrimbutGerbil and @DrBlackbird , just to make sure that @JustineMumsnet and the legals are aware.

So I don't know about you, but it feels a bit crime-adjacent to me.

Isn't that what 'forensic' means? "Relating to or denoting the application of scientific methods and techniques to the investigation of crime".

Further to @DrBlackbird 's comment, this is something I have been meaning to do for a while: a list of all the other collections on the Aston Database (FoLD). It's, um, quite enlightening

Actually I might not do them all, but here are the first few...

Perverted Justice Chatlogs - Full archive of chatlogs from now defunct site perverted-justice.com depicting online instant messaging interactions between convicted child sex offenders and adults posing as minors.
Closing arguments in US rape trials
Corpus of German Suicide Notes
Meredith Kercher case file library - from murder case
Transcritpions from Kurdish and Arabic questionnaires
Emergency call transcripts of innocent and guilty callers - A collection of plain-text transcriptions of 60 emergency calls. 30 were placed by innocent callers (i.e. witnesses) and 30 by deceptive callers (i.e. the perpetrator pretending to be innocent).
Romance Fraud Messages
Transcripts of Laughter used in Police Interviews
Deceptive Opinion Spam Corpus v1.4 - odd Tripadvisor reviews
Complete collection of the Unabomber writings
Transcript of oral proceedings in the Court of Appeal (Criminal Division) brought by the ’Stansted 15’ (Thacker et al), November 2020
The Italian Red Brigades statements during the kidnapping of Aldo Moro
Racist tweets and countermessages by Stop Hate UK
Corpus of Contemporary English Legal Decisions, 1950–2021
Catalogue of abusive language training data
The Shipman Inquiry
Pledges to Harm
Threat letters transcribed data from the FBI
Michelle Carter/Conrad Roy text messages - from a manslaughter trial
Operation Heron corpus - The dataset is a subset of the abusive letters sent by Margaret Walkers to individuals in the public eye between 2007 and 2009

That's the first page. The next page includes cyber-hate, trolling and child sexual abuse. So I don't know about you, but it feels a bit crime-adjacent to me.

https://134.151.36.51/recent-submissions?offset=0

Sidenote, this url has been set up to be difficult to acccess... It was not like that a few weeks ago.

MN’s data analysis will be / would’ve been in interesting company being part of that conference. Deviant behaviour. Fake news. Murder trials. Suicide notes. Child sexual abuse. False emergency reporting. Pick up artists.

We are/were to be included along with a session on tackling online offensive and hateful language. Good to know where a forum for mothers and threads by women seeking to protect women’s rights and safeguard children’s stands in the eyes of Aston’s FL institute. Slow handclap to NM.