I’m looking for answers from instance admins, if you’re a regular user, you can still answer but it’s more helpful for me to get answers directly from admins.
If a user on [instance A] asked another instance (Instance B) to remove their federated account and federated content copies from instance B (likely also banning it so content doesn’t continue to flow) would the user on Instance A be in trouble with their instance admin for asking for such a thing.
Obviously it depends on the instance’s rules but that’s part of why I’m asking the question, to get answers from instance admins on this.
On one hand I can see how it would since, since it hurts interoperability and can create tension between instances, but on the other hand a user has the right to be in specific places or not be in those places, that probably extends to not wanting to be federated into an instance they find objectionable (assuming it is for good reasons).
Not an admin, but from a legal perspective, users in the EU have the right to request deletion of their data under the GDPR, which the consequences of violation are up to €10m or 2% of annual turnover (not profit), whichever is higher
Frankly, if a user asks a service owner to delete their personal data, the service owner should do it as promptly as possible.
Lemmy doesn’t federate “personal data” to other servers. The GDPR has a strict definition what can be “personal data”. The Wikipedia has a good overview of the relevant laws in various countries: https://en.wikipedia.org/wiki/Personal_data
Requesting the deletion of posts and comments that they agreed to be federated when signing up is purely voluntary but usually done as it is fairly easy to ban a user and delete their contributions.
From your link
The “directly or indirectly” part is important here, a username is a constant identifier between a user’s posts and comments
Given comments and posts are free text input, there’s no way of knowing the entire set of a user’s content doesn’t contain PII, unless an admin wants to spend the time combing through and determining which posts definitely contain PII and which definitely don’t, they should delete it all. The data subject does not need to make specific listings of what they want deleted, the onus is on the service owner to be able to process the deletion request completely and within a timely manner.
No, as only the instance admin that hosts the original account can indirectly associate a user handle with actual “personal data”. An admin of a federated instance can not, as they do not have any “personal data” to correlate it with.
If a user themselves posts “personal data” publicly it is not covered by the GDPR (IANAL) and thus not subject to mandatory deletion requests. Of course deleting everything is often the easiest course of action, but this is not legally required.
Also not a lawyer but I’ve done a lot of GDPR training since it was introduced and I believe you’re incorrect—the data subject posting it publicly or not doesn’t factor into the validity of a deletion request under the GDPR. There are a limited set of specific reasons a service owner can refuse a deletion request and they’re pretty much down to preventing abuse and facilitating compliance with other laws.
Not a lawyer, but honestly, both of these takes are probably not correct.
I’d say that most fedi-services fall more into the ‘can I make someone delete an email’ GDPR category (tldr: probably not, but maybe) with a dose of the ‘this service is for personal/non-commercial use and includes messaging and social media’ exemption.
This of course won’t work if you’re taking money or doing commercial activity but at that point you’re a business and should consult your lawyers to ensure your compliance. (And if you can’t, then maybe don’t be in that business.)
I wouldn’t want to be the one to spend the billion dollars to litigate that, but frankly if you’re not in the EU, and not a business, then the person demanding removal would have to take you to court to force compliance (assuming you didn’t just do it so you don’t have to deal with a grumpy person) which is… unlikely.
The much more horrifying interpretation is that the data controller, processor, and sub-processor language comes into effect and everyone needs to sign written agreements with every other fediserver to be even remotely in compliance.
Except that only applies to federated servers that exist in the EU. If your data gets federated out to a country outside of the EU, they don’t have to listen to your whines of GDPR as it’s not enforceable. And given that you could be federated with hundreds of instances across the world, good luck.
I said the same thing with AI scraping. All someone needs is to add their own instance that federates with everyone else and they can scrape data for AI training till their heart’s content.
https://gdpr.eu/companies-outside-of-europe/
https://gdpr.eu/what-is-gdpr/
Cool cool, now realistically, do you have the time, resources and know how to find and contact every owner of every federated instance these comments have made to? Would you be able to deal with the legal resources of any number of jurisdictions to truly test whether that is actually enforceable?
My point basically is that it’s functionally impossible regardless of what the law says, and you should treat your comments and personal information as such that they won’t ever be able to be deleted or scrubbed.