Mastodawn

Hawkwinter :firefish:

So - I saw @[email protected] has a

[QUOTE]My posts are licensed under Creative Commons Attribution-NonCommercial-ShareAlike (https://creativecommons.org/licenses/by-nc-sa/4.0/)[/QUOTE]

And it got me thinking: It would be cool if you could license your posts out under terms for-profit corporations wouldn't like (IE Viral copyleft licenses), and make them want to block / defederate your account, and not want to use your art or images in an AI model (such as because they would be legally required to publish their software sourcecode to the world under a CC-BY-SA license or AGPL) - and then if they wanted to do so, a class action lawsuit could make them regret it.

I dunno if CC-BY-SA or AGPL would get us there. But it would be cool.

Any Fedi-Copyright Attorneys have any suggestions on getting there?

Creative Commons — Attribution-NonCommercial-ShareAlike 4.0 International — CC BY-NC-SA 4.0

Richad A Jul 21, 2023

@Hawkwinter @[email protected] I was asking about this a few weeks ago. :-). I like the idea that the Fediverse is CC by A for non commercial use.

Kathrin Jul 21, 2023

@Hawkwinter @aral

I am not an attorney, this is not legal advice.

It's a *very* slippery slope. Currently the entire fediverse works on a good-faith basis, where we don't worry about the copyrights of posts. Once that kind of mindset arrives here, it's going to lead to a lot of destruction.

For one: Federated servers continuously copy from each other, simply due to how the technology works. Servers have to cache posts from other servers locally, even though nobody gave them permission.

1/3

Kathrin Jul 21, 2023

@Hawkwinter @aral

It would be 𝘦𝘯𝘵𝘪𝘳𝘦𝘭𝘺 too easy for a slew of copyright trolls to come along, make an account on a tainted server (perhaps one of their own), create an account that easily can go viral with original copyrighted works, wait for the works to proliferate through the fediverse, and then sue any server that has copies of the work.

The only things that make this a little inconvenient are the different jurisdictions servers are in.

2/3

Kathrin Jul 21, 2023

@Hawkwinter @aral

Currently, we just don't have the numbers and the server operators are too poor to make this viable, but it's ripe for the picking.

What would be required to stave this off would be a unified TOS that most servers would incorporate that would allow that kind of sharing. And that, in turn, would require more skilled copyright attorneys to work it out. Perhaps the EFF might help in that regard.

3/3

Hawkwinter :firefish:Jul 21, 2023

@Kathrin @aral

I see what you mean. A compatible ToS check baked into the federation handshake to determine if and how the servers communicat, and setting the terms at the server level rather than individual users.

I also see what you mean about it being something that would be hard to do without something like the EFF.

Fedi-Servers cachingWhy DO fediservers cache all the posts? Why not just save the sources, and then when the user goes to read the posts, load them from the original sources in the browser in Javascript?

Kathrin Jul 21, 2023

@Hawkwinter @aral

Those are good questions and I'm not nearly qualified to answer them. It would be worthwhile to get some input from fediverse developers on these, though.

If we leave it up to the individual users to license their posts, it'll never get the widespread adoption it needs. Also, Creative Commons is not a good license for it, because it always requires attribution AND a mention of the license itself (in every work, i.e. post), which is not feasible.

Ketil Froyn Jul 22, 2023

@Hawkwinter @aral @Kathrin

> Why not just save the sources, and then when the user goes to read the posts, load them from the original sources in the browser in Javascript?

I think you would find this to be slow/unstable in practice. And if a post is boosted a lot, or as it hits the federation timeline on busy sites, the number of hits from viewers across the fediverse could probably take down a small instance. In that sense, this would probably raise the bar for hosting an instance.

Ketil Froyn Jul 22, 2023

@Hawkwinter @aral @Kathrin and a malicious site could federate a cool message, and serve an evil message directly to all the users later.

Hawkwinter :firefish:Jul 22, 2023

@kefir @aral @Kathrin

and a malicious site could federate a cool message, and serve an evil message directly to all the users later.Can't they already do this by editing a post after it's boosted a whole bunch, and it would just take a bit longer to get updated?

Ketil Froyn Jul 22, 2023

@Hawkwinter @aral @Kathrin I guess so, but it would still be noticed and potentially filtered or flagged on all other server when it is federated.

Hawkwinter :firefish:Jul 21, 2023

@Kathrin @aral

Hmm.

People here might be operating on a good-faith basis, but you know that Threads and AI companies scraping people's data to monetize it are not.

The idea is to make it so the people operating in good faith unaffected, while people operating in bad faith are incentivized to not scrape the data, with meaningful consequences for if they choose to do so anyways.

Kathrin Jul 21, 2023

@Hawkwinter @aral

You are correct, but this is a political issue. The problem we're facing is a legal one. Threads and AI companies have the money and legal teams that we do not.

Like it or not, we are small and powerless. We need to plan accordingly.

ck Jul 21, 2023

@Hawkwinter @aral
I think if the GDPR has taught us anything then it's that for-profit corporations tend to not give a fuck about the legality of their actions or your personal rights.

Hawkwinter :firefish:Jul 21, 2023

@ck @aral

Oh. I'm assuming as much. The question is can it be made such that there are real consequences when they don't, and they have to start giving a fuck about your rights.

ck Jul 21, 2023

@Hawkwinter @aral
Sure, we just have to convince the majority of people to vote for representatives that honestly care more about people that about corporations so that they may pass laws accordingly.

Matt Boehm Jul 21, 2023

@aral s we can set any terms we'd like. "Use of this content in a machine learning dataset is prohibited"

The Open Buddhist University Jul 22, 2023

@Hawkwinter @aral Not a lawyer but worked in this area for a few years on the tech side: current law in the US allows for processing and limited repurposing of copyrighted material by algorithms. This is the same fair use exception used by e.g. Google Search to scrape the web and present search results which may include snippets of the e.g. article. Most tech companies will, however, voluntarily respect robots.txt files, so if you don't want your site scraped, add one! https://www.cloudflare.com/learning/bots/what-is-robots-txt/