Yesterday over 700 creators, actors, songwriters, musicians and authors launched the campaign „Stealing isn’t innovation“ to protest against AI companies stealing their intellectual property to train their AI models: https://www.stealingisntinnovation.com
Why this is necessary you might wonder?
Well the big public LLMs are trained heavily on copyright protected data without asking the creators/owners for consent. Remember the flood of Ghibli-style images last year? ChatGPT was only able to create this style because it was training on the work of Studio Ghibli - without any compensation for them. Another example: we know that Meta used LibGen, one of the biggest pirate collection of books and research papers (it includes illegal copied 7,5 million books and 80 million research papers) to train their Llama model. The theft of intellectual property is not an accident, it happens on purpose.
So, what happened here: Gen AI creates value based on the creative work of humans - without compensating them. It it part of the business model to use every data and work that is technically available without asking for permission or consent.
Do you remember the time when Napster and later torrent became popular? It was made crystal clear that copying and sharing music and movies were illegal (technically the sharing, but since it was a substantial part of the process, everybody who downloaded is sharing as well). There were even campaigns stating that piracy of music and movies is a crime and bring you in jail for up to 3 years. Piracy was (and in this case still is a crime). So, what is the difference to what big tech is doing with the creative work of humans at the moment?
We all should be much more aware of what it means to use this technology and that there is an enormous external price - societal economical and ecological - that needs to be paid.
And a little fun fact at the end: what would happen if no more humans create things, big tech can use to train their models? There is research showing that if models that are trained on AI-created things only or mainly, degrade and collapse.. The technology does not function without the human input…
On a little side note to this campaign: I really like it, I think it is super important. Only thing that make me a little sad that the initiators speak about the us creative industry only. It’s not only there. Big tech is stealing creative work from humans around the globe.
Sources:
The campaign: https://www.stealingisntinnovation.com
Article about Meta and LibGen: https://www.wired.com/story/new-documents-unredacted-meta-copyright-ai-lawsuit/
Research about training AI on recursively created data: https://www.nature.com/articles/s41586-024-07566-y
Thorsten Jonas is the Founder of the SUX Network and Digital Sustainability, Sustainable UX and Responsible AI consultant, Keynote Speaker and Author. He is guiding teams and companies in crafting sustainable, responsible and ethical digital products, advocating for a sustainable digital world and co-leads as invited expert the UX Chapter of the W3C Web Sustainability Guidelines.
