With the expanding use of generative AI tools like Midjourney, Stable Diffusion and DALL-E in the creative industry also comes looming legal and ethical concerns. Generative AI is trained on large datasets created by human artists, an unethical process that involves scraping copyrighted material from the Internet without consent or compensation. This article argues how, contrary to the claims that it ‘democratises’ art, generative AI has the opposite effect of automating and privatising the creative process, taking protections away from individual artists, while also failing to replicate the artistic cognition that makes creativity morally valuable. However, how copyright law has already been evolving to catch up with AI’s development by defining how AI art can be considered theft. From the US Copyright Office’s decisions on AI creations Zarya of the Dawn and A Recent Entrance to Paradise, which highlighted the importance of human authorship; to the European Court’s 2017 ruling that Pirate Bay was infringing copyright by encouraging users to make copies of protected works; and finally to the lawsuits against Stability AI and OpenAI, what the law has shown is that it does not afford the same rights to machines as it does to humans.