The admin of the Mastodon instance cyberspace.social just received an AI powered notice to delete the parody account @microsoft@lea.pet
Let’s save this post for the next follow friday…
thanks microsoft! i just followed the account!
…yet.
Microsoft is such a horrid company.
Microsoft is at this point made up entirely of technically inept people running fraudulent mechanical Turks.
After forced to switch to Microsoft teams, I became their PM and developers are alien origin.
Like, some of the logic doesn’t make any human sense. Like, LLM slop would’ve done better.
$followers++This is honestly a dumb post. It doesn’t say anything about “Microsoft” understanding or not understanding anything.
It just shows them using an automated system to try and take down an account that they think is infringing on their trademark. There are legal protections for parody accounts, but they are not absolute and it’s possible that Microsoft could get a court order compelling the owner to cease control of the account.
So I guess I’ll explain the post for you, since it seems you do not understand the fediverse, either.
cyberplace.social federates with lea.pet but is not responsible for and cannot control the Microsoft parody account on lea.pet. Microsoft sent the notice to the wrong server. That’s why Microsoft does not understand the fediverse or how federation works, and why the post is funny instead of dumb. Now you can laugh too! You’re welcome!
Holy shit. I can’t believe that 65 people upvoted this. Do you really believe that’s how the world works? Are you actually adults?
Lol please explain how the world works
You have full control over a server on which you chose to run certain Software. But you feel you don’t have to comply with takedown requests because that’s just how the software works? That may work on indulgent parents but not in court. If you’re too technically inept to know how to comply, then you’re just not complying. End of story.
LOL you think a court will enforce this C&D for a parody account??? 🤣 are you an adult??
So it’s funny because the fediverse is so niche that no one designing automated copyright systems care about its odd and unique addressing system?
An attempt at censorship failed because the censors didn’t understand the system they were trying to censor. I think that’s both funny and satisfying.
They didn’t care about this system. It just got caught up in their news sources.
This isn’t funny, it’s just a thing that happened.
☝️: 😐
Yes, exactly! Now you get it!
:D
Not that I have any sympathy for them but technically speaking, can we say “Microsoft does X” when it was just a “brand protection LLM”?
For sure they chose it to be an outward connection but are we really now stating that an LLM can represent a company? I feel like that’s both “you are stupid for using LLMs that way” and empowering LLMs at the same timeThe phrasing is really weird too. Like I would understand if they were asserting their trademark, but saying it’s an account they’ve lost access to?
They don’t even say they lost access. Lol they just say delete it. Fuck these assholes.
They say they do not have access to it.
Yeah. I don’t have access to it either. What does that mean? Of course they don’t have access to it. It’s not their account.
They are the ones who configured the tool to be used this way. Yes, Microsoft did that.
You’re always responsible for the actions of those that you choose to represent you. Regardless of how stupid they may be.
The instance rules state;
3 . No impersonation of a person or a brand. Even parodies.
6 . Each username on here should be a person, not a brand or corporation - and that person must be you, no impersonation.
Now I’m not coming to the defense of a corporation because fuck corps but wouldn’t this account violate those rules? Microsoft is a (shitty) brand and this is a parody of that.
I might be wrong but thats how I would interpret those two rules.
And again fuck microsoft, even if it does violate the instance rules it should be left alone just to piss of a shitty corp.
I believe the Microsoft account is hosted on Lea.pet, which doesn’t have those rules
don’t harass people if there isn’t a valid reason for doing that
what a nothing rule lol, every harasser believes they are justified
This rule is my favorite lol
I like rule 8
Poor Ubuntu.
What does DNI mean?
Do Not Interact (with us) lol
:D
Some might argue it requires a terminal level of brain worms to understand what is being said with rule 8. But I for one find beauty in this exactly level and flavor of petty, and I come to the fediverse for it.
Also, what’s a valid reason for harassment?
Presumably using Ubuntu is one of them lol
Fair point, I didn’t know those rules weren’t on lea.pet. That’s on me I should have checked, my bad.
Mrrrrp :3
pats
Quick, everyone make a million parody accounts to troll microsoft harder.
could be a rule break of number 9
are we sure on the English translation?
@Una@europe.pub pls translate
“Reject humanity emprace meow meow mrrrp :3”
Ah I see, thanks ☺ Good kitty
meow meow :3
You’re making the same mistake Microsoft made
It’s not hosted on that instance. It’s hosted at another instance. You’re making the same mistake that the Microsoft bot AI did.
Is it a mistake? Wouldn’t federated content still count the same way legally, since an instance is also a website?
you’re assuming that American law applies to people not living in America.
I’m not assuming that, I just don’t see why would it even matter if it’s from another instance.
Yeah, I literally made an edit admitting it so thanks for wasting your time pointing it out again.
Stupid rules deserve breaking.
6 . Each username on here should be a person, not a brand or corporation - and that person must be you, no impersonation.
“I’m not a businessman, I’m a business, man!”
-Jay-Z regretting his words
enf.microsoft
Do they know what enf means on the internet?
Microsoft doesn’t understand parody
So, here’s what I would do. I would comply (you should be able to delete the local instance of that account). But I’d also reply pointing out that it’s a mirror of the real account hosted at lea.pet and their real beef is with them, and should that user interact with or generate content pushed to you, the local copy would be re-created.
Keep a copy of the email you send (because it’s highly likely a human doesn’t monitor that mailbox) and then move on with your life. If a real person then wants to complain you can just forward the email you sent and tell them the same still applies.
It’s automated and the email indicates as such.
Why comply? As far as I’m aware there’s no legal obligation to do so. They think they can just ask for things and get them. Fuck them.
Well, legally there’s no reason to comply. At the same time I personally have no skin in the game and deleting the account locally won’t do much (unless you purge their content too).
Sure, but why even do that minor thing for them? Just ignore it like it deserves.
When a website links to another like Facebook or similar is the website not at least partially responsible for content, like if the content was terrorism they might ultimately blame Facebook or whoever but still expect the website to remove the content from their page?
Sure, but there’s no obligation to give a company the username with their company name.
So, then, according to corpo logic, you’ve setup a system that automatically repeatedly breaks the law (copyright , maybe? Ianal).
And if the liability is on you for hosting a federation service with no control over the content and you are accountable for the replicated content, then it is effectively the end of federation. Again, not a lawyer but given how much the law favors corps over individual rights in the US it seems like it would track.
I’ve said this before. The UK online safety act if they enforce it hard against fediverse instances, it will be the end of federation, for UK users without a VPN at least. Because it puts too much on the shoulders of small site operators.
In this case though, the exception most countries have for site operators to avoid being responsible for their user’s posts is usually reliant on action being taken when content on your site is reported to you. There isn’t really an exception for saying “Umm, wasn’t from my site mate. Go follow the trail and get the original guy”. The argument will be, the site you control has the content, remove it.
In the UK in the 1990s there was a court case [1] that might even form the part of the case law behind the publisher exception. In that case the claimant stated that the ISP was alerted to forged usenet articles (usenet was pretty much a good analogue for modern federated content) that he believed defamed him. They did not remove the articles (presumably because they did not originate on their usenet server, by their users I am not sure). He sued them and the court ruled in his favour. There’s more nuance, but the take away is pretty much what we got in the law created later.
Since then we have enacted the Defamation Act 2013 [2], which has section 5 that gives SOME exemption to operators of websites that allow posts by third parties (that pretty much covers the fediverse). That makes it clear that if the claimant cannot identify the user (which would be the case for 99% of threadiverse users), and if you are informed about the content and do not take action, then you may be held accountable for the defamation. Now that just means that if they tell you X post is defamatory or should be removed for another legal reason, if you refuse to do so in a reasonable time period, then you can be held responsible and treated as the publisher of that message. So if it were to breach some law, they could sue you for it as if you posted it yourself. Which is kinda why I’d say just remove it if there’s any doubt at all. I’m not a legal expert, but that’s how I read the act.
I’m not sure how it works elsewhere. I live in the UK. But generally the rules are somewhat similar.
[1] https://en.wikipedia.org/wiki/Godfrey_v_Demon_Internet_Service [2] https://www.legislation.gov.uk/ukpga/2013/26/section/5
Very interesting.
I’m making a decentralised open source sharing protocol (have your free cloud drive, website, …) and I very often gets the question about what if my node shares some illegal stuff (because when someone shares yours, you share theirs)?
It’s all encrypted so a node cannot know what it shares, and if someone asks you to take down abc.xyz then we’ll do it and it should be the end of the story.
Seems that’s what it would be in the UK at least (and the rest of the EU usually doesn’t have harsher laws for what I know), what are your thought?
On the one-hand I think it would be similar to how usenet works now for binaries. That is, once notified under DMCA (for the USA) and likely similar laws in other western countries you’re duty bound to remove it.
I don’t know if there would be other problems with hosting files when you don’t know what they are. Also in terms of defence against DMCA, how would the original file uploader defend against it when you can’t know what the file is without the key. Person A reports file xyz as infringing their copyright, Uploader B says it doesn’t. Normally you could re-instate it and let the two parties fight it out in court. But, I wonder how it would play out when you hosting the file don’t even know what it is.
I’m really not sure how it would really stack up against copyright law in general and more specifically laws for truly illegal content (e.g. CSAM), since you could be hosting that and never know.
Seems a bit more of a risky venture to me and more a question for an actually qualified legal advisor I’m afraid.
Maybe look up the story behind Tornado Cash.
EU law has liability exemption for hosters if they remove illegal content once notified (DSA Article 6). I think we still have to wait for more case law to know how that works with encryption. National law may still cause problems.
Thank you!
Seems like the EU Safe Harbor Provisions, you basically must not incentivice illegal hosting, accept takedown requests, but also have some sort of procedure for the takedown requests. Which all seem quite easy to follow and adhere to and would function perfectly for Tenfingers IMO.
Toronto cash seems to avoid the “not incentivice illegal usage”, and after a quick check it seems to be almost it’s sole reason to be, please correct me if I’m mistaken here.
Toronto cash seems to avoid the “not incentivice illegal usage”, and after a quick check it seems to be almost it’s sole reason to be, please correct me if I’m mistaken here.
I’m not familiar with the details. The point is simply, don’t expect to get away with playing games.
Seems like the EU Safe Harbor Provisions, you basically must not incentivice illegal hosting, accept takedown requests, but also have some sort of procedure for the takedown requests. Which all seem quite easy to follow and adhere to and would function perfectly for Tenfingers IMO.
Yes, but there is more. The DSA is written in a very convoluted way, with the exceptions for smaller platforms scattered here and there. I don’t remember what exactly applies here. You may also have obligations under the DMA, CRA, and quite probably the GDPR.
I feel like this is going to become a problem with federation in the future. A Mastodon instance is hosting content outside of its control that may or may not comply with its internal policies or local law. Is that instance protected legally? Likely not.
It would likely be treated the same way as auto forwarding an email would be treated.
AT proto’s PDS architecture does solve that 👀
Has BlueSky implemented federation yet?
No, but if it did you could control where your data lived, because people do host their own PDS iirc for bluesky.
wdym they not have federation yet, there are independent servers for every service (e.g. pds, relay, appview, moderation) except for PLC
Yeah, if.
yes
I don’t know. You may be generally already protected to some degrees by existing laws for users interacting with your site. This time it’s just users interacting with your site via another server. If you defederate from bad instances and have a good DMCA system, you might be okay.
Although, this type of thing could be one of those things that’ll take a court case or updated legislation to solve.
IANAL, but most law that I’ve heard of regarding third party content requires the site hosting the content to conform to takedown notices issued. So, having a good DCMA system requires you to be able to take down content from instances that may not be bad, but governed differently.
As for the law “catching up with” federation sites, I don’t see that happening unless Mastodon and Lemmy start creating massive lobbying arms.
As for the law “catching up with” federation sites, I don’t see that happening unless Mastodon and Lemmy start creating massive lobbying arms.
Forgot that’s how America worked.
In the UK, you could just keep making noise until a member of parliament gets interested and mentions it once
So how does UK law handle federation?
No clue. Online Safety Act is a bin fire
Maybe you should make some noise until an MP cleans up the issue.
When did they understand anything? Fuck MS
this is why niche stuff on fedi is so funny sometimes. especially when a corporate entity tried to approach it, in the most corporate way lol.
