4 ms·

Honestly, I’d like the UK to go heavily in favour of data mining. I think a lot of copyright holders are just trying a money grab because AI is a rapidly growin

Honestly, I’d like the UK to go heavily in favour of data mining. I think a lot of copyright holders are just trying a money grab because AI is a rapidly growing industry

artninja1988 2y ago
Yes, I agree. I essentially conveyed the same message in my email comment to them. Feel free to do the same
cycomanic 2y ago
So the AI industry is not a money grab essentially training their models on all the work of other people while keeping their own "sauce" secret? If we allow AI companies to disregard copyright then they should be required to publish their models. You can't have your cake and eat it too.
Silhouette 2y ago
I think a good starting point is that economic incentives should support most the people who are contributing the most. If we allow training AI models to become a method for laundering the exploitation of others' creative work without informed consent or compensation - and let's not pretend that this isn't exactly what has been happening - then I think we will start to see the same problems we've already encountered with search engines that display the most important parts of the original content directly on their own pages but potentially at a much larger scale. The result is similar in each case. It becomes economically and/or emotionally unsatisfying to make the effort to create new and better works. If commercial organisations that just rehash all that hard work for their own benefit without giving anything back will be the ones who ultimately benefit the most from it then sooner or later they may find no-one wants to gather or share the data they'd like to train their next generation of models on any more. We'll just see more and more paywalls with contractual terms for access gating more and more of the information on the Internet. This kind of exploitation without compensation is exactly what copyright is supposed to prevent. It's how it forms an economic market for creative work that functions in a familiar way. It is astonishing to me that governments have become so caught up in the AI hype of the past couple of years that they are falling over themselves to undermine their entire creative industries for fear of missing out. I fear this will not end well when the AI bubble bursts and we aren't all welcoming our new AGI overlords in 2026 because it turned out that the maths was still essentially the same as it had been for a long time and there wasn't another quantum leap in the scale or quality of available training data waiting for the AI companies to exploit.
frereubu 2y ago
Two wrongs don't make a right. There are certainly corporate interests who try to exploit copyright law to the maximum, and I agree that that should be stopped by reducing copyright terms etc, but that doesn't mean that, for example, the authors of books should be exploited by both AI companies and their publishers just because you don't like the way publishers exploit copyright. There's more at stake than one type of company vs another.

Honestly, I’d like the UK to go heavily in favour of data mining. I think a lot of copyright holders are just trying a money grab because AI is a rapidly growing industry

artninja1988 2y ago

Yes, I agree. I essentially conveyed the same message in my email comment to them. Feel free to do the same

cycomanic 2y ago

So the AI industry is not a money grab essentially training their models on all the work of other people while keeping their own "sauce" secret? If we allow AI companies to disregard copyright then they should be required to publish their models. You can't have your cake and eat it too.

Silhouette 2y ago

I think a good starting point is that economic incentives should support most the people who are contributing the most. If we allow training AI models to become a method for laundering the exploitation of others' creative work without informed consent or compensation - and let's not pretend that this isn't exactly what has been happening - then I think we will start to see the same problems we've already encountered with search engines that display the most important parts of the original content directly on their own pages but potentially at a much larger scale. The result is similar in each case. It becomes economically and/or emotionally unsatisfying to make the effort to create new and better works. If commercial organisations that just rehash all that hard work for their own benefit without giving anything back will be the ones who ultimately benefit the most from it then sooner or later they may find no-one wants to gather or share the data they'd like to train their next generation of models on any more. We'll just see more and more paywalls with contractual terms for access gating more and more of the information on the Internet. This kind of exploitation without compensation is exactly what copyright is supposed to prevent. It's how it forms an economic market for creative work that functions in a familiar way. It is astonishing to me that governments have become so caught up in the AI hype of the past couple of years that they are falling over themselves to undermine their entire creative industries for fear of missing out. I fear this will not end well when the AI bubble bursts and we aren't all welcoming our new AGI overlords in 2026 because it turned out that the maths was still essentially the same as it had been for a long time and there wasn't another quantum leap in the scale or quality of available training data waiting for the AI companies to exploit.

frereubu 2y ago

Two wrongs don't make a right. There are certainly corporate interests who try to exploit copyright law to the maximum, and I agree that that should be stopped by reducing copyright terms etc, but that doesn't mean that, for example, the authors of books should be exploited by both AI companies and their publishers just because you don't like the way publishers exploit copyright. There's more at stake than one type of company vs another.