UK Online Safety Regulations and impact on Forums

Tested the AI illegal harms addon with a couple of posts including a photo of a gun and it worked very well. Reported and sent for moderator approval.
 

Attachments

  • AI Test 2.webp
    AI Test 2.webp
    4.9 KB · Views: 12
  • AI Test 8.webp
    AI Test 8.webp
    20.8 KB · Views: 16
@Alvin63 Are you using the standard version of the AI Bots addon, or did you go for the upgrade with OpenRouter support?
No idea. It's an addon I got from genesis that uses Open AI. It's very good - maybe doesn't do absolutely everything I want. But it's content moderation as opposed to link scanning it seems so maybe need something else for link scanning. Or maybe need an upgrade.
 
No idea. It's an addon I got from genesis that uses Open AI. It's very good - maybe doesn't do absolutely everything I want.

I was looking at this update for it, so you can choose to use other AI models with it, rather than just ChatGPT.
 
That looks good. The standard Open AI works very well for content moderation though. I guess it maybe needs something else to scan links.
 

I was looking at this update for it, so you can choose to use other AI models with it, rather than just ChatGPT.
AI bots sounds good though!
 
Keeping it just to use for illegal harms content though or it could get too expensive. I've spent 7 dollars so far with about 15 tests. There are other options in XF to send certain things for manual moderation.
 
Just thinking about the ability to route requests through OpenRouter to use my Google API Studio key and the free Gemini 2.5 model:
I'd upgrade just to get OpenRouter's ability to configure fallback AI LLM models in case one fails to respond. Don't use that addon, but using OpenRouter with ability to configure multiple AI LLM models as fallback and configure fallback providers and even the order in which you fallback is awesome to have :)
 
Could someone explain all that please :-) What exactly does it do that's awesome?
see
For facial age analysis data, being able to configure to only use models that do not retain data is nice too https://openrouter.ai/docs/features...uiring-providers-to-comply-with-data-policies. I have data_collection set to deny :)
You can restrict requests only to providers that comply with your data policies using the data_collection field.


  • allow: (default) allow providers which store user data non-transiently and may train on it
  • deny: use only providers which do not collect user data
 
Back
Top Bottom