UK Online Safety Regulations and impact on Forums

None are foolproof unless constantly done. So if you are checking your users constantly (and I've seen suggestions that laptop/mobile cameras make this an "excellent" idea) all you are proving is that someone at the time of testing passed the checks (yah! Big Brother really will be watching you!).

If the "point in time" check(s) are anonymous then whoever "passed" the check has little to worry about, if it's a case of borrowing an adult's ID, etc then I suspect the fallback would land on the owner of the ID should there ever be any fallback. Is it now "my" responsibility to stop the kid pinching their parents phone whilst unlocked and posting or viewing something on facebook and ending up "harmed" by it, or should the parent have the responsibility. Although in that rather contrived situation how would anyone even know?

Realistically the only less intrusive method I can think of for constantly validating age would be using an LLM to do analysis on writing styles for each user and have an estimate of age generated off that - then based off that do periodic more reliable checks. That might be realistic for a large social media site that has that kind of compute, but I think not for the smaller players. So unless anyone is going to go properly OTT on age checking it's always going to be at best just another factor for consideration and ultimately how you "run the place" is really what will matter.

I guess I will spend the weekend reading the guidance and decide between:
  • Age gating the entire forum (so it's just adults or adults at time of checking!)
  • Age gating some features of the forum as risk reduction method (private messaging really)
  • Just tightening up the rules on the occasional more colourful prose or double entendre and sticking with 18+ tickboxes (to at least indicate to kids it's not for them)
 
Last edited:
Alvin63 said:
Yet Ofcom says if you have age verification you don’t need to go through a lengthy child risk assessment. So Id rather just meet the requirements for age verification.
Click to expand...
But this can impact a site's traffic as you can see from linked articles, sites that implemented age verification saw their traffic drop as visitors just move to find another site to visit. You really would need die-hard visitor/members who would go through age verification route!
 
A quick look at their examples of PC and PCC makes me think we have little to fear about in our forums, They are all basically things we don't allow anyway and delete as soon as they are reported. So we are mitigating that risk as required.

Conversation/PMs are a different matter but they can be restricted to people we can verify the age of. Its bad news, but doesn't mean we have to totally cripple the forums or age-verify every new signup.
 
Just to save anyone ploughing through. The headline content Ofcom lists is:
  • Primary priority content
    1. Pornographic content
    2. Suicide and self harm content
    3. Eating disorder content
  • Priority content
    1. Abuse and hate content
    2. Bullying content
    3. Violent content
    4. Harmful substances content
    5. Dangerous stunts and challenges content
  • Non-designated content
    1. Depression content
    2. Body stigma content
Plus any other non-designated content you can think of I guess ... although there was a bit about needing to tell Ofcom if you think of something (not fully read yet). Hopefully in reading it will become clear quite what this stuff will be - I mean if you ran a Model Railway Modelling forum would it be something around encouraging users to use tools like knives and drills and files which could cause injury without proper PPE for instance, or is it is something else?
 
eva2000 said:
But this can impact a site's traffic as you can see from linked articles, sites that implemented age verification saw their traffic drop as visitors just move to find another site to visit. You really would need die-hard visitor/members who would go through age verification route!
Click to expand...
I realise it's a step but that's why I was looking for something simple and easy - like a selfie or email check. I know my members wouldn't mind because they are desperate to get the forum back! It may put some new members off. I think the banking age check one would put people off, even though it's probably the most reliable.

I think the article was referring more to dodgy sites perhaps? Unless I misread.
 
chillibear said:
Just to save anyone ploughing through the headline content Ofcom lists is:
  • Primary priority content
    1. Pornographic content
    2. Suicide and self harm content
    3. Eating disorder content
  • Priority content
    1. Abuse and hate content
    2. Bullying content
    3. Violent content
    4. Harmful substances content
    5. Dangerous stunts and challenges content
  • Non-designated content
    1. Depression content
    2. Body stigma content
Plus any other non-designated content you can think of I guess ... although there was a bit about needing to tell Ofcom if you think of something (not fully read yet). Hopefully in reading it will become clear quite what this stuff will be - I mean if you ran a Model Railway Modelling forum would it be something around encouraging users to use tools like knives and drills and files which could cause injury without proper PPE for instance, or is it is something else?
Click to expand...
Thanks. So not much different than in their previous documentation. So have they left "injured animals" off the list now then for Priority content? The main issue on forums maybe is the bullying content label as it's a grey area when people disagree on something.

The bit I posted earlier is one of the big sticking points - about swift moderation to take something down.
 
Sorry to ask you, I should look myself but had a hectic day. But any mention of links to other sites again? Technically youtube should be age verifying anyone who clicks on a youtube link.......
 
eva2000 said:
Latest version of my facial age verification system with frontend MVP demo (left image) and dedicated test endpoint to test 2d portrait photo images and various age thresholds (18, 21, 23, 25 etc) against my AI facial age verification analysis to see how it performs for age estimation, confidence level, and its reasoning for such determination. It has been surprisingly accurate for 2d portrait photo images, at least within 0-3 years of known age, using a small set of tested images. But there have been some false results as well for folks who do not look their age.


View attachment 321798View attachment 321799
Click to expand...
Just tried this version. Worked fine on computer but I noticed the Cloudflare thing didn't work automatically - the box needed ticking first.
 
This is the bit I found interesting

PCU C2 Have a content moderation function that allows for swift action on content harmful to children, where it is currently technically feasible to take appropriate action

What is the point of that? If it's a nod to smaller forums who may not have enough moderating power, then surely it's also a get-out for bigger forums so doesn't really protected anyone! It's contradictory. "Swift action " "Where it is currently technically feasible" - what does that last bit even mean.

I think I'll keep focusing on age verification - I don't want to go through hundreds of pages of more noddy documents........
 
Facial age recognition not that popular though

www.internetmatters.org

Age assurance and online safety: What parents and children have to say

Age assurance views from parents and children | Internet Matters. See parents' and children's views on age verification online.
www.internetmatters.org www.internetmatters.org

Screenshot_20250425_232924_Brave.webp
The top concerns parents and children have with age assurance methods are about privacy (43% of parents; 31% of children) and how their data will be used (35% of parents; 30% of children).

Overall, children are less likely to have concerns about age verification but are more likely to be worried about access to documentation (18%). This is driven by 9-12 year olds who responded with this, compared to 13-17 year olds (21% cf. 15%). This is likely driven by fear of being locked out of services altogether.

A third of parents (34%) are concerned that children could get around age verification methods.
Click to expand...
Screenshot_20250425_233522_Brave.webp
 

New UK Internet Policing Law Targets US Online Forums | ZeroHedge

ZeroHedge - On a long enough timeline, the survival rate for everyone drops to zero
www.zerohedge.com www.zerohedge.com

Online forums based in the United States that rely on First Amendment protections are getting caught up in internet regulations in the UK, where they now risk being blocked under recent legislation.
Click to expand...
Gab, an American social media network, positions itself as a champion of free speech.

Gab CEO Andrew Torba said in a March 26 social media post that the UK government has demanded that it submit to “their new censorship regime under the UK Online Safety Act.”

Gab—which has no legal presence in the UK—was informed in a letter from UK regulator Ofcom on March 16 that it falls specifically within the scope of the law and must comply.

Under the OSA, sites that allow user interaction, including forums, must have completed an illegal harm risk assessment by March 16 and submitted it to Ofcom by March 31.

Ofcom warned that noncompliance could result in enforcement action—including massive fines of 18 million pounds (more than $23 million), or 10 percent of a company’s annual revenue—or even court orders to block access in the UK.
Click to expand...
 
The most popular option seems to be parent or guardian proving age, but I can't see how that works as how would they prove it? I would be age verifying adults anyway :-)

So I'm trialling Shufti Pro. Their backend seems a bit slow to load. Facial estimation seems to work well. Tried ID Document as well and that works well and you can opt to only extract date of birth from it. The first one I tested (myself) came up with a tick to show it was verified. After that it kept coming up with a message about waiting for approval? Did you find that @chillibear? Even though it showed as approved in the backend.

The only thing is, that before the user starts, they need to tick a box saying they are over 16 and consent to Shufti using and storing their data and biometrics. I think that will put people off.
 
You must log in or register to reply here.
Back
Top Bottom