It’s emerging public knowledge that #AICompanies are going to have to pay for #TrainingData. I’m assuming that this will happen.
Given that. Should #Medium participate on behalf of our Authors and how should we pass that money on to authors? The per article price is not going to be very much money, say $0.10. But we could put the money into the author payment pool and pay out by Quality/Popularity.
For people saying No. How do you feel about your current licensing? Medium's default is "all rights reserved" to you.
But it's not proven yet, and not at all clear legally, whether that still allows AI companies to consume your content under a legal theory of fair use for derivative works.
@coachtony It would need to be opt out for me. Either for the entire account or on an individual story basis (like checking the box for meter or email when publishing)
For those who opt in, I imagine they'd rather have the 10 cents
The option for opting out would be required for many who will not participate in handing over free content for others to earn from.
@coachtony Excuse my ignorance, because I'm a little confused. Are you saying that Medium is looking into giving content created via AI more prominence? If that's the case, what will happen to content crafted by human brains?
@cendrineartist no, this has nothing to do with what you read on Medium. AI companies base their models on having read nearly every article on the internet. The question for all platforms is whether we want to continue to allow that.
@coachtony Ok, then, why would Medium need to get on the bandwagon? You aren’t an AI company. Unless you want to keep an eye on content that infringes copyright?
@cendrineartist @coachtony AI companies train their models on data they get from the internet. People are doing their work and later AI company will take it for free, teach their model and later earn billions on it.
So people are not happy about this and want AI companies to pay for the data. Because the data authors don't get anything for having an AI company take their data.
That's all.
@djvdq @coachtony Of course people are not happy about it. And that’s totally understandable. Many have tried to take advantage of technology before for a quick buck.
It’s not the first time creators are not paid for their work. I know first hand.
My question then is: What will Medium do against it? Because they allow it as long as people mention it in their posts. Are there any plans to monitor it more closely?
@coachtony If I catch an AI using my stuff, I am dragging their ass to court and I shall be more furious than a 7-headed mythical dragon.
@attilavago go for it. they have already done it
@coachtony They have? How? 99% of my articles are behind paywall? Is Medium allowing for bots to parse content?
The comedy here is that AI has been trained by us all for free for many years.
Who played "chase the Pokemon' Ball around town " 15 yrs ago?
@coachtony Is the idea that the "legitimate" AI companies will pay up, and the unscrupulous ones will just steal it anyway? So might as well get paid for it?
Honestly, I'd opt out whether they can still access it or not. Getting paid for the data legitimizes their use of it. If it's just scraped from the internet, at least I can still say that it was done without my consent.
@coachtony a formal disclaimer at the end should clear this out from the author perspective. Then, give the authors the option to disclose an article for training (similar to metering it under the MPP).
Options are required because many will not type on Medium or elsewhere for free.
@coachtony The times change very fast in this business. AI is in its infancy, but there are many very, very smart people working hard on it.
Intellectual property law is complex and slow-to-change. Your motivation as a chief executive officer is different from mine as a writer, but there is connection. We already do not get paid much for our work, and I am confident you will do the best you can when Big Money comes a-calling.
Where will that "BIG MONEY" land?
@coachtony Licensing for AI training data should work the same way as it does for music. For every listen, the artist gets royalties. Every single time content from my article (even if just a sentence) get surfaced by AI, the AI company needs to pay me royalties. Any other model is exploitative.
@coachtony I say no. Or make it an option to opt out.
Agreed. The option needs to be required. Power mongers are out of control.
Too many do not understand the way they cut the nose off their own faces.
@coachtony I’d opt in, and wouldn’t want to get paid for it. But I’m not a big author trying to make my money publishing
If Medium does participate, it should be opt-in, with a *very* clear explanation of what it entails both e-mailed to all members and posted to the blog. Maybe even a Q/A video session with you and AI experts? There's a *lot* of confusion about this burgeoning topic.
@coachtony
Giving writers the option sounds reasonable. Instead of opting everyone in without their consent or knowledge, let's have it as a setting writers can toggle on. Promote it so writers know about it and understand the rewards and tradeoffs of licensing their writing for LLMs. Also could it be anonymized somehow, in case someone doesn't want #ChatGPT to be able to write a story in their style?
#Medium #AI #LLM #TrainingData #AICompanies
@coachtony I've been freelancing as a functional tester for years now, hence this new Gen-AI training approach seems tantalizing for Medium authors
@coachtony I'd really rather not have my content used to train AI models.
The climate around AI is, already, fraught, and the conversations around its uses do not default to nuanced, unfortunately.
As much as the "it's just another tool" set want their view/sentiment to be true, history has shown us the ways technology and tools inevitably become what the majority want them to be. Or even "need" them to be.
@coachtony Well, I think the best option would be to let the author decide. Some authors don't like that their content is used to train AIs. Others don't care.
How to redistribute the possible payment?
It depends on how AI companies pay.
If they pay per word, pay authors per word. And so on.
@coachtony My vote is no. #Medium should not sell authors' words on 'behalf of authors.'
We've had many changes to platform recently--AI! AI pubs! AI images! and as an editor, I'm culling (boring ass crap) AI writing out of the submissions queue. As a writer, I don't want my words used as training grounds. AI will use it anyway, so why not make money? That argument suggests that Google should also be able to sell our email addresses and passwords, as the hackers will get them anyway.
@coachtony Not a fan of opening the door to AI scraping from Medium. Writers are nervous about AI to begin with and now allowing our creative souls to be plagiarized for 10cents per article is a punch to the gut. If Medium decided to do that writers should be given the option of opting out through a checked box.
@darrenweir it’s already open
@coachtony I get that but by accepting money for it aren’t we encouraging it and saying that what they are doing is okay? I know there’s no stopping it now but I don’t have to support it.
@coachtony Count me as another vote for yes, but with an opt-in. Someone else made an interesting point about quality/popularity not being relevant for AI training, but I still think you'd have to take that into account for distributing the payment pool--otherwise users would be incentivized to just post as much spam as possible.
@coachtony If the question is whether AI companies do it for free or at least pay up since our writing is going to be used anyway, whether we like it or not, then yes, sure. I'd still like some form of opt-in/out in case that becomes relevant. Not sure how payment based on popularity would work... so can't add an educated comment on it. I suppose, if you do decide to go that route, do what is best for all writers and not a select few whose popularity do not necessarily reflect value or utility.
@coachtony
Medium and all of AI can and should need to pay an author for their work and obtain permission to use it.
No business gets materials for operation or use without payment of that material.
The mop and the front door cost cash.
Granting free use of material that is not owned by the grantor is purely wrong.
Someone give me free stuff to sell to others for my gain please... or maybe no "please" is needed- Yikes!
FYI, a @medium article referring to this thread showed up in my e-mail daily digest and web page "For you" tab, due to you having highlighted it. I don't follow this author. As posts about Medium aren't supposed to be eligible for distribution, this seems to be a loophole. (Indeed, I am still seeing "Meta posts" like this all the time.)
https://medium.com/new-writers-welcome/medium-wants-to-sell-our-content-for-10-915b83c01bc9
https://help.medium.com/hc/en-us/articles/360006362473#distribution-disqualification-0-3
@coachtony Do it and I'm out.
@hollyjahangiri does it matter to you that they already took it?
@coachtony and you should know that from the efforts I expended to get ichi(.)pro shut down (at least they're smart enough they stopped scraping my articles - there is SOME payoff for the effort).
@coachtony let me qualify that: the price paid is too low and individual contributors should be able to opt in or out.
@coachtony I'm assuming you're saying that they did - you know that for a fact?
@hollyjahangiri yeah. Most of the internet. I think there was a WashPo article about all of the sites.
@coachtony Google did it with books, years ago. Some of us fought it, but the precedent helps all the scrapers.
@coachtony Since our copyrighted work is most likely already being used with AI it would be nice to see some of it compensated back to the author of the work.
@coachtony No. Not only no, but FUCKING NO!!!!!
@coachtony Should Medium participate? Not unless every member is given the opportunity to OPT OUT.