SoylentNews is people
SoylentNews
Arthur T Knackerbracket has processed the following story:
[...] When the UK government launched a public consultation on AI and copyright in early 2025, it likely didn't expect to receive a near-unanimous dressing-down. But of the roughly 10,000 responses submitted through its official “Citizen Space” platform, just 3% supported the government's preferred policy for regulating how AI uses copyrighted material for training. A massive 88% backed a stricter approach focused on rights-holders.
The survey asked for opinions on four possible routes the UK might take to address what rules should apply when AI developers train their models on books, songs, art, and other copyrighted works. The government’s favored route was labeled Option 3 and offered a compromise where AI developers had a default right to use copyrighted material as long as they disclosed what they used, and offered a way for those with the rights to the material to opt out. But most who responded disagreed.
Option 3 received the least support. Even the “do nothing” option of just leaving the law vague and inconsistent polled better. More people would prefer no reform at all than accept the government's suggestion. That level of disapproval is hard to spin.
It's a triumph for the campaign by writers’ unions, music industry groups, visual artists, and game developers seeking exactly this result. They spent months warning about a future where creative work becomes free fuel for unlicensed AI engines.
The artists argued that the fight was over consent as much as royalties. They argued that having creative work swept up into a training dataset without permission means the damage is done, even if you can opt out months later. And they pointed out that the UK’s copyright laws weren’t built for AI. Copyright in the UK is automatic, not registered, which is great for flexibility, but tough for any enforcement, as there's no central database of copyright ownership.
Officials crafted Option 3 to try to appease all sides. The government's stated aim was to stimulate AI innovation while still respecting creators. A transparent opt-out mechanism would let developers build useful models while giving artists a way to refuse. But it ultimately felt to many creators like all the burden fell on them, and they would have to constantly monitor how their work is used, sometimes across borders, languages, and platforms they’ve never heard of.
That's likely why 88% of respondents went for requiring licenses for everything as their preferred choice. If an AI model were to be implemented, wanting to train on your book, your voice, your illustration, or your photography, it would have to ask, and potentially pay first.
A final report and economic impact assessment from the government is due in March. It will evaluate the legal, commercial, and cultural implications of each option. Officials say they will consider input from creators, tech firms, small businesses, and other stakeholders. Clearly the government's hope to smoothly start implementing its prefeerred appraoch won't happen.
For now, the confusing status quo remains. Without a court ruling or legislative fix, uncertainty reigns. AI developers don’t know what’s allowed. Creators don’t know what’s protected. Everyone's waiting for clarity that keeps getting delayed.
What happens next could shape the UK's digital economy for years. If officials side with the 3% who backed their initial plan, they risk alienating the very creators whose work is so valuable. But stronger licensing rules would undoubtedly face resistance from AI startups and international tech firms. Either way, the fighting is far from over.
(Score: 2) by VLM on Thursday February 05, @04:28PM (1 child)
As a kid I used a "learn to draw" book to make poor quality drawings of cartoon woody woodpecker(tm). Some decades later I like to imagine I could still churn out really bad woody woodpecker(tm) drawings.
Copyright should not stop creative people from generating something creative from a copyrighted inspiration.
People who know nothing about copyright law think that if I see a picture of a dandelion flower in a copyrighted gardening book, then I draw a funny meme of a dandelion flower that vaguely resembles the pic I saw a long time ago, somehow I personally am violating the copyright of the gardening book. Not so, as a creative interpretation my drawing is copyright me, today, not the gardening book from years ago.
What DOES apply, and people often are confused by, is trademark law. I can't sell hand drawings of woody woodpecker(tm) identified as "a woody woodpecker(tm) drawing". I could sell drawings that don't infringe, maybe an IRL woodpecker bird thats not anthropomorphized like WW.
"most" of the arguments against AI That are reported by people who know nothing about the law as "copyright" problems are actually "trademark" problems. If you pay money for someone to provide you a drawing or video of Superman(tm) and Superwoman(tm) "greatly enjoying each others company in a biologically reproductive sense" that is a major trademark violation expect to be sued.
If you ask a rando human to "use your years of reading shitty fantasy novels to create an imaginary NPC to meet at the Inn in our next DnD session" that is 100% legal. I read tons of copyrighted works to create that creative work of art, but that doesn't matter. As long as I'm not dumb enough to rip off a trademarked character like, say, Frodo(tm) from LotR(tm), I'm safe.
AI should be about the same. If I can read copyrighted material, including for free at the public library...
The danger of going down the path of anything copyrighted "deserves" infinite rent-extraction is it can and will be applied to humans. Every time I as a human solve a simple math problem, say, calculating a percentage tip at a restaurant in my head, how much money should I be required to send not one, but every, copyrighted elementary school textbook I was forced to read in elementary school? If I swear on a bible in a courtroom that I never read the textbook before doing the homework, which in some cases is probably even true, do I still have to pay the copyright tax on something I learned independent of the copyright holder decades ago? If I quote "Romeo and Juliette" verbally in private to another person, which 1980s copyrighted printed version from my middle school English class, do I owe the money to for violating the book's copyright by making that quote? Will I be permitted to legally prove I read a copy of the play where the copyright expired decades ago (it's pretty old...)?
The government can and should come down like a stack of bricks on trademark violations, especially commercial trademark violations, but I don't see any good outcome for copyright law abuse as an anti-AI weapon.
Another good topic is plagiarism, and abuse of plagiarism laws. If as a human I quote and properly attribute the front page of the New York Times as containing an article with the title "Research Finds Interaction With Father, Not Mother, Affects Child Health" (from the NYT website for Feb 5 2026) that is just a factually accurate and correctly cited quotation in a discussion of child health or father's rights or whatever. Anti-AI activists think they will "fix" everything by requiring payment every time a properly cited quotation is made, but I think that will F things up for humans a lot more than it'll F things up for AI.
(Score: 2) by Freeman on Thursday February 05, @05:59PM
https://gutenberg.org/ebooks/1513 [gutenberg.org]
Romeo and Juliet was probably always public domain in the United States. Still, public domain books/etc. haven't really been a thing the average user cared about all that much until the advent of computers. (Except for cost of cheaper reprints.) Computers all of a sudden allowed the average person to have a giant library of books on your computer without taking up the physical space having 1,000 or even 100,000 or 1,000,000 books would require.
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"