UK proposes letting tech firms use copyrighted work to train AI

Or straight from the horse's mouth without having to navigate the Grauniad's attempt at mining your personal data: https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence  :-)

Sounds like the US tech firms have bent Starmers ears and he is allowing them to use your work unless you make the effort to opt out. Goes completely against copyright law as it exists and companies like ChatGP have been effectively “stealing” people’s work. 
it’s an even bigger joke they are talking about AI generated images being given Copyright protection only the idiots in Whitehall could come up with this. 

I presume this means train the AI to promote Government agendas too. Will read the papers in detail tomorrow.

It reads like the government are trying to force greater transparency from the A.I tech companies, especially if they are using rights holders work without their knowledge. Then ensuring that the rights holder is appropriately compensated for that use after agreeing such use. Copyright for Computer Generated work is already protected in the UK.

This was written by a human being with the aid of a computer and keyboard, revolutionary idea I know.

I think it's quite a well-reasoned consultation that seeks to address both sides of the issue.  In short, they suggest three options:

1. No nothing;
2. Opt-in (AI companies can only use data if the rights holder has given express permission to do so);
3. Opt-out (AI companies can use any data unless the rights holder has expressly refused permission for them to do so)

They are suggesting option 3 as the best way forward. In addition, there would be some form of compensation for the use of the rights holders data and/or greater transparency over what data has been used.

If anyone has particular views I suggest they respond directly to the consultation?

Edited by Gothic Image

I agree with tandi and Gothic Image - at least to some extent. This is an area where I had expected the British Government would pursue a policy of having no policy, preferring instead to allow the courts to sort out the mess. So top marks to our Government for at least trying to protect us this time. However, I fear this is where the good news ends.

For a start, it's doubtful that anyone in government (or in the opposition) has the technical knowledge to understand the raft of complex technical issues and the nuances contained therein. Instead they will rely upon advice from consultants. No doubt these consultants will be largely from the same LFITCs (large foreign information technology corporations) that already enjoy significant patronage from the British Government.

Meantime the legislation itself will likely be characteristically British - botched, hurried, badly framed, watered-down and generally half-arsed. Consequently, the courts will still have to sort out the mess. And, as we all know, the LFITCs have much deeper pockets than we do. Therefore it is reasonable to expect that most of the subsequent legal challenges are unlikely to be resolved in the public's favour.

I would add that in my view, legislation of this sort, at this point in time is far too little, far too late. Government is attempting to shut the stable door, while the proverbial horses are galloping off into the sunset, several km away.

Fact is, LFITCs already have our data, masses of it, and it's already been assimilated, at least to some extent. IMHO, the rot set in over two decades ago when LFITCs such as Google and Yahoo pulled an immense sleight of hand, conning the public into allowing them to "cache" our images in their "search engines". Thus the majority of the planet's pictures are already "in the system".

And, FWIW Gothic Image I will indeed be responding to the consultation. :-)

Edited by waist.it

An update from Petapixel with a slightly misleading title: https://petapixel.com/2024/12/20/uk-photographers-reject-government-plan-to-allow-ai-companies-to-train-on-their-data/

Gothic Image thanks for the link. Yes it is a tad misleading.lol :-)

FWIW, I'm busy formulating my response to the consultation. In doing so, it would seem to me that Government is in a position to make a very important additional condition to use of AI in this manner. Namely that all "knowledge" gathered should be open sourced and made a matter of public record - so that everybody can use it.

I don't think I am alone in my aversion to allowing foreign billionaires to steal our intellectual property and make themselves even richer than they are already, at our expense. However, if one removes the profit motive and makes whatever is gathered a matter of public record, then I think much of the current skulduggery would pretty much evaporate. After all, why go to all the effort of stealing something if one has to give it all back after it has been stolen?

More importantly, AI in this context would become a true publicly-owned resource that we can all enjoy. It wouldn't stop AI being developed. In fact I suspect the lads and lasses at Stable Diffusion would love the idea. But it would go a long way to ensuring AI was used for the good of the many rather than the profit of a few.  Granted, I'm struggling to frame that last concept in a manner that yer average politician might actually understand. ;-)

Fortunately the consultation doesn't have to be completed until 2025-02-25. Moreover, consultation site allows one to save one's answers and come back to them later. I was quite impressed with this feature actually. If one is required to give thoughtful responses, one needs time to pause for thought. :-)

Anyway, a link if anyone else wants to have their say...

• https://ipoconsultations.citizenspace.com/ipo/consultation-on-copyright-and-ai/consultation/