Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Japan’s government will not enforce copyrights on data used in AI training

technomancers.ai

Post reply on HN
Print view View original

141–150 of 426 posts

Re: Japan’s government will not enforce copyrights on data used in AI training

#141

Post by sclarisse »

shagie wrote: post #34
scottiebarnes wrote: post #12

So if you train an audio model on say, Eminem's voice, then write some songs and have it perform them...Would this output be legal to publish?

I've been amused by the WH40k videos narrated by David Attenborough. There was some discussion about this a few years ago regarding Lyrebird - https://news.ycombinator.com/item?id=14182580 In particular, celebrities have an additional right - Right of Publicity. https://www.law.cornell.edu/wex/publicity > In the United States, the right of publicity is largely protected by state common or statutory law. Only about ha…

> I've been amused by the WH40k videos narrated by David Attenborough.

Have you seen the Thomas the Tank Engine videos narrated by Ringo Starr and George Carlin?

…

Re: Japan’s government will not enforce copyrights on data used in AI training

#142

Post by GaggiX »

lamontcg wrote: post #132
nl wrote: post #89

Earlier quoted context omitted.

I strongly agree with this. There's a distinction between "learning from" and "copying". "Learning from" is a transformative process that distills from the observation. This distillation can be as simple as indexing for a search engine, or as complex as a deep neural network. Simply because a neural network can create something that is a copyright violation doesn't mean the training process itself it. A human can see…

> There's a distinction between "learning from" and "copying". Neural nets can memorize their training data. Generally that isn't what you want, and you strive to eliminate it. However, it could instead be encouraged to happen if someone wanted to exploit this law in order to abuse copyrights.

The law applies to the training of a neural network; you're not depriving the copyright holder of his intellectual property; if you use a copy of his work, he still owns the copyright independently if you copy it by right-clicking > copying or by overfitting a generative model.

Re: Japan’s government will not enforce copyrights on data used in AI training

#143

Post by __loam »

fnordpiglet wrote: post #19

I think this should generally be true. The aggregation performed by model training is highly lossy and the model itself is a derived work at worst and is certainly fair use. It may produce stuff that violates copyright, but the way you use or distribute the product of the model that can violate copyright. Making it write code that’s a clone of copyright code or making it make pictures with copy right imagery in it or…

The collection of copyright works for the explicit purpose of processing them for a for-profit ML model has not been shown to be fair use, and the fact that many are being marketed as for profit products that meaningfully compete with the original works is a strike against them being fair use.

Re: Japan’s government will not enforce copyrights on data used in AI training

#144

Post by kelnos »

carom wrote: post #80
mannyv wrote: post #22

Earlier quoted context omitted.

There's no difference between an art student looking through a museum or archives for ideas and an AI using the material for training. Same could be said for reading. A medical student reading through textbooks or a writer who reads is essentially what an AI is doing. You can ask an art student to create something in a certain style. You can get writes to write in a certain style. Equivalent.

Conflating training a model with human learning is wrong. When training a model you are deriving a function that takes some input and produces an output. The issue with copyright and licensing here is that a copy is made and reproduced numerous times when training. The model is not walking around a museum where it is an authorized viewing. It is not a being learning a skill. It is a function. The further issue is tha…

I'm glad you brought this up, as this tendency for people to anthropomorphize a learning algorithm really bothers me. The model training process is a mathematical function. It is not a human engaging in thought processes or forming memories. Attempting to equate the two feels wrong to me, and trying to use the comparison in arguments like this just feels irrelevant and invalid.

Re: Japan’s government will not enforce copyrights on data used in AI training

#145

Post by pcurve »

nl wrote: post #113

> With the effective implementation of AI, it could potentially boost the nation’s GDP by 50% or more in a short time. Err.. No, it won't. That's a ridiculous, laughable statement. Japan 2022 GDP: $4.1 Trillion Amazon 2022 Revenue: $513B Google 2022 Revenue: $279B Microsoft 2022 Revenue: $198B So even growing a brand new Amazon, Google and Microsoft in "a short period" would be insufficient to grow GDP by 50%

yeah. There are countries growing at 7% annually, but they're mostly in Africa. Niger, Rwanda, Congo, etc.

But please for the love of God, don't compare GDP to revenue.

Re: Japan’s government will not enforce copyrights on data used in AI training

#146

Post by it_citizen »

nl wrote: post #113

> With the effective implementation of AI, it could potentially boost the nation’s GDP by 50% or more in a short time. Err.. No, it won't. That's a ridiculous, laughable statement. Japan 2022 GDP: $4.1 Trillion Amazon 2022 Revenue: $513B Google 2022 Revenue: $279B Microsoft 2022 Revenue: $198B So even growing a brand new Amazon, Google and Microsoft in "a short period" would be insufficient to grow GDP by 50%

Comparing the revenues of tech companies to the gdp of a country, even to give a sense of scale is comparing apples to oranges.

Even if a bit unlikely, I would not be completely surprised if the service industry as a whole produced twice the value its produces today thanks to AI in the next 20/30 years. Not to mention the productivity gains in other sectors.

Re: Japan’s government will not enforce copyrights on data used in AI training

#147

Post by kazinator »

swfsql wrote: post #136
kazinator wrote: post #78

What's surprising here is that Japan is usually crazy gung ho on copyright enforcement ... at least against individuals. So it's kind of disgusting to see this relaxation, when it suits some corporate or national interests. https://en.wikipedia.org/wiki/File_sharing_in_Japan "Unlike most other countries, filesharing copyrighted content is not just a civil offense, but a criminal one, with penalties of up to ten years…

I see the other way around, their usual crazy gung ho on copyright enforcement against individuals is disgusting. This new stance is marvellous and examplifies how free people should interact.

Yes, that too; but a double standard on top of that is double plus ungood disgusting.

Re: Japan’s government will not enforce copyrights on data used in AI training

#148

Post by Dylan16807 »

kelnos wrote: post #138
fnordpiglet wrote: post #19

I think this should generally be true. The aggregation performed by model training is highly lossy and the model itself is a derived work at worst and is certainly fair use. It may produce stuff that violates copyright, but the way you use or distribute the product of the model that can violate copyright. Making it write code that’s a clone of copyright code or making it make pictures with copy right imagery in it or…

> The aggregation performed by model training is highly lossy and the model itself is a derived work at worst and is certainly fair use. Lossy or not, the training data provides value . If all the various someones had not spent time making all the stuff that ends up as training data, then the model it trains would not exist. If you are going to use someone else's work in order to make something that you are going to…

> the training data provides value

I wish it was easier to build on someone else's copyrighted value. Geforce Now shouldn't have to get permission from game makers to rent out server time for users to play games they already own. Aereo shouldn't have been legally obliterated for the way they rented out DVRs with antennas. Using a sample in a song shouldn't be automatic infringement.

Re: Japan’s government will not enforce copyrights on data used in AI training

#149

Post by greentext »

If I compress an artist's painting into a jpeg and rehost part of it for individual t-shirt designs I am committing a crime.

If I compress an artist's painting into a model & rehost what's essentially a highly flexible complete version of their painting for infinite, perpetual use of any kind ... I'm not committing a crime?

Re: Japan’s government will not enforce copyrights on data used in AI training

#150

Post by kmod »

fnordpiglet wrote: post #131
kmod wrote: post #114

Earlier quoted context omitted.

This is the "guns don't kill people, people do" argument. Not saying that that proves things one way or another, just that assigning responsibility is not really a cut and dry question and many people think that it's important to look prior to the final interaction IANAL but I believe in the US tools that are designed to circumvent copyright are illegal, which makes sense to me inasmuch as one believes that copyright…

Except these models are not designed to circumvent copyright, in fact their primary purpose is to generate non-copyright and non-copyrightable output. It can be induced to product facsimiles of copyright material, but that’s explicitly not the purpose or intent and requires positive action on the users behalf to occur.

I'm saying `A => B` (something should be illegal if it's primarily for crime) and you're saying `-A` (LLMs are not primarily for crime), which is not really a disagreement. My point is disagreeing with the GGP who argued that `-B regardless of A` (LLMs should not be a crime regardless of whether they facilitate crime, because it is the end-user who is committing the crime).

I happen to mostly believe the conclusion (`-B`) but not the particular argument.

Post reply on HN