I thought that mumsnet now screened all pictures before being allowed to post???

[LoudSnoringDog]

How has a man been able to post his penis on one of forces sweethearts posts??

thought after the child abuse pictures a few months ago this shit wouldn’t be possible again??

Whichever way you look at it, security just isn't good enough if this is still happening. And clearly, using AI because it's cheaper than actually employing someone, isn't effective either!

As these disgusting images always seem to appear overnight UK time and there is no paid moderation during those hours, maybe the next step is to ban images from being posted between 2300-0700.

Its strange because every time I post a recipe or a mundane picture of a pair of shoes or a top, it takes ages to be posted.

It is if they're flagged for review @MsDDxx - possibly our least favourite job!

@BeckyAMumsnet

What do you mean ‘if’ they’re flagged for review? I don’t think a basic AI screening filter is going to be good enough anymore, particularly regarding the Online Safety Act.

Websites facilitating or enabling cyberflashing (the act of sending unsolicited sexual images) are now illegal in the UK and it carries serious consequences?
Cyberflashing is a criminal offense under Section 66A of the Sexual Offences Act 2003, as inserted by the Online Safety Act 2023. I imagine ‘enabling’ could include not applying any/or sufficient protection to prevent it.

This isn’t directed specifically at you BeckyA because you are, presumably, an employee but, can you please make sure that someone responds to this - in this thread? I am going to report my own post so it reaches someone - please don’t bother replying directly to me on my email as I don’t actually access that email connected to this account anymore - please simply reply in this thread.

I have to add that, once again, I’m surprised and disappointed at the ‘oops silly us’ tone to responses from MNHQ regarding this issue.

Hi @ParvuliThankYouDebbie Just to clarify, our moderation system includes multiple layers: user reports, human moderators, nightwatch , and AI-supported flagging. We use one of the most advanced tools currently available, but as with any automated system, it’s not perfect, which is why we don't rely on it in isolation and we're actively strengthening our approach.

We appreciate that you’ve taken the time to raise this. but also want to note that if you interpreted any of our responses as flippant today, that certainly wasn't our intention.

Not really good enough. Time for paid moderators.