I've been doing fine with Google workspace and no issues for just under a year. But then my forum isn't that big (about 750 members). I have a weekly mail goes out to everyone but only if they meet certain criteria So it's actually only to about 200 people each week.
Thanks. I solved it anyway - it was our IP address. Tried rebooting first to try and get a new IP address and that sorted it. But - both IP addresses had blacklist marks! So won't be surprised if it happens again. I don't even use a vpn which apparently is one cause of it.
I've been getting this a lot online recently, only to find it has now happened on my own forum with embedded videos. There is nothing to actually click on to sign in! If I click to take me to youtube, where I'm already signed in I still get it. I actually have youtube videos set to not link...
Well Chat GPT just wrote me a 2026 addendum covering the Women and Girls risks and the two new priority harms. So seems to cover it! I'm trying Claude as well and see which one I prefer.
Edit: Amazing. The Claude AI one was extremely thorough and comprehensive. Just the odd bit needs...
So did you put in the 2026 guidance and put in your current risk assessment and get ClaudeAI to revise it? And did you then have to read through it all to check ClaudeAI got it right? (Just trying to work out if it would take longer to check than to do :-))
Re the women and girls requirements. The document mentions four main harms covered:
Online misogyny
Pile-ons and online harrassment
Online domestic abuse
Image-based sexual abuse.
I think this is quite easy to cover. For me I can just add those caterogies to the AI content moderator and say...
So - for some reason I always think Google AI is female and Chat GPT male 😂 Just comes across that way. Anyone else think they seem gendered and what about other AI's?
It's a legal requirement for everyone with a forum to review and update the risk assessment in line with current legal requirements. As @chillibear said, I think the 9 points about women and girls have to be incorporated in the existing references to priority harms. Or something like that.
In the past year there has been nothing illegal for the AI content moderator to catch. However the AI content moderator has had lots of false positives some of which are hilarious. On pet forums people call their pets boy or girl sometimes - and talk about cages. So plenty of false positives...
Oh more as well then! About women and girls - that isn't mentioned in the article I linked about (which also gives info on what people missed off or got wrong on risk assessments). CSEA and CMAS are now four different categories apparently.
I think it might either need a new record keeping...
Nothing has really changed on my site so I was planning to just do a short document saying the risk assessment had been reviewed and no changes. However, there are now apparently two/three new illegal harms you have to do the risk assessment for as well. So I'm thinking of just doing those...
Anyone done their second year's risk assessment yet? Mine isn't due until May as I closed down in March last year and re-opened in May. Did you fill out a whole new risk assessment form or something else? Just wondering how to do it when nothing has changed. Thanks.