Last month, video editing software CapCut quietly joined Adobe Firefly and Meta in updating their terms to enable AI scraping: a process that extracts creators’ content to train generative models. The 500 million creators who use CapCut now have no control over when their likeness, voice, or videos will be taken to generate AI content.
These changes extend beyond account deletion, as the new terms grant CapCut perpetual rights to sublicense any content previously made on the platform. Although some artists have taken action and filed lawsuits against certain AI platforms for data harvesting for AI art, current outdated copyright laws do little to protect artists from this new craze.
Artistic erosion and theft
This erasing of artists is already visible across many creative industries. Greg Rutkowski, a Polish digital artist, has had users enter his name as a prompt on Stable Diffusion, an open-source AI art generator, over 93,000 times, making him one of the platform’s most common prompts. Artists like Rutkowski are neither credited nor compensated, even though AI models mimic their signature styles without consent.
It’s not just visual artists, either. The generative craze is sweeping up novels to produce fan fiction in an author’s writing style, while AI music models mimic artists’ voices and songwriting. Though some users may enjoy these adapted pieces, the artists inspiring these flash-in-the-pan outputs are paying the price.
Can’t creators simply say no?
No, they can’t.
Avoiding AI in today’s creative industries is nearly impossible. 89% of marketers globally rely on Meta platforms for business promotion. Yet Meta uses publicly shared content to train its AI systems, which means that Creators can’t opt out when the very tools they need to build a career double as pipelines feeding AI models. And when these platforms do offer opt-out options, their processes are often buried within pages of terms and conditions.
What’s the point in writing a novel if AI can throw together a story that will guarantee social media success? Why would a consumer commission an artist if an open-source AI art generator can replicate their style for a fraction of the cost? How can artists protect themselves if the tech platforms that were once designed to support their creative endeavours are now undermining their work?
Removing the humanity from creativity
The impact of AI scraping will have a knock-on effect on consumers, too. AI models can imitate a painter’s brushstrokes or an author’s syntax, but they cannot replicate the years of lived experience and creative process behind the work. As AI continues to develop, audiences risk losing the ability to distinguish human artistry from algorithmic mimicry.
Generative AI web tool, Stability AI, scraped over 5.8 billion images to train its models, while another, Books3, consumed at least 170,000 books without consent, including the works of many living authors.
This unprecedented scale is not isolated to a few artists, but spans multiple creative industries. Since ‘style’ is not protected under copyright laws, AI scraping is almost impossible to regulate. Without a clear framework, artists are left exposed to their creative work being hollowed out without permission or compensation.
Looking ahead: what is the future for AI art?
There is no denying that AI has become the hottest buzzword of the decade so far. With the number of tech corporations in London raising by 26% in 2024 alone, this recent surge feels similar to the dot.com boom — a period of rapid technological and internet growth in the 1990s and 2000s. Though the internet revolutionised modern society, the initial website-making craze died down soon after its peak.
Perhaps AI will follow this trend as its novelty wears off. Or, perhaps, new methods of combating AI scraping, like Nightshade — a digital tool that invisibly alters artists’ work to poison AI training models — will become mainstream.
However, the damage AI scraping has already done to artists’ careers, and to the value of human originality, cannot be undone.
