and lack of transparency*

And yes, I understand that we need age verification to protect minors, ID Verification and Face Scans just aren't the way to do it due to it's data privacy concerns a dn lack of transparency like this.

I'd be interested to see a lawyers comments on Discord and Veratads policies around this.

Also, to anyone that says "Oh just don't do it", some people kind of have to.
Lets say you moderate a discord server, and someone sends content that gets blurred by discord, we need to be able to verify that the user posted something innapropriate ourselves before taking action at risk of mod abuse.

I guess it depends on if a Government ID counts as "sensitive personal information"
The severe lack of transparency here is deeply concerning, expecially if this gets extended globally instead of just the UK and Australia.

From what I can tell, although I'm not great at reading legal policies much, the Veratad privacy policy does not state if said information is deleted or not.
If anyone wants to take a better look at it and correct me they can, I could only find "we do not store sensitive personal information"

Additionally, they don't explicitly mention that the data and photos are deleted immediately, unlike say Google or Roblox who do specify this during the process, I had to lookup Discords article for this which has it in the F&Q section ( support.discord.com/hc/en-us/art... )

Or even change your password to something random like keyboard spam and then never try to gain access to it again.
Stop crossposting, or even checking on twitter sure, just make sure to maintain the handle to prevent impersonation.

For some people, especially those with a large following, "staying on Twitter" is slightly essential.
If they close their accounts, theres a chance someone could make an account in their name and attempt to impersonate them.

I think the correct thing is more "Private your account, and log out."

And it won't take long before a client modification like Vencord has a plugin to hide these features client-side.

My point is that the "protection effect" people want is already in place.
If you don't authorise the AI app, you are opting into nothing.
As for the server component, yes, discord does need to implement permissions that do not allow these to function to protect images in servers.

The screenshot I sent was this process, which is required before being able to use any of the 4 AI apps.
The prompt appears on attempting to use one of the AI apps, and cancelling it means you cannot use the AI apps. (Refusing does not affect discord.)

Have you tried the process yourself?
OP's post was how server owners cannot disable it, which is correct.
In terms of the policies, before being able to use any apps, including these AI ones, you have to authorise it and agree to the policies.

I'm not trying to defend Discord here, but these are opt-in services.
Before being allowed to use the AI apps, you have to authorise them, which is you agreeing to their TOS and Privacy Policy and giving the AI app a bunch of permissions.
You do NOT have to agree to the apps policies to use Discord.

I believe (take this as a grain of salt) when you authorise an app (which you have to do to use any of these AI services, or just any discord applications in general) you are also agreeing to the services privacy policy and terms of service.
This happens with all apps, like Dyno, YAGPDB, Etc.

policies*

The best way to "opt-out" is to not even test the services, by authorising the app, you are effectively opting into the service.

Unfortunately because these are external apps as opposed to an internal feature, I don't believe discord will add a permission to block this for server owners.

It's also interesting to read through the privacy policites / Terms of Services for these 4 AI's that for the most part think any content they create is legally theirs.

Been testing this out and can confirm the following:
- Disabling External Apps / application Commands / Activities does not disable the feature (Only limits it to being seen by the user running it).
- The AI's can respond with 18+ content in non explicit channels, particularly with "Roast".