Generative artificial intelligence ("Al") has increasingly become a focal point in legal discussions, raising complex issues across multiple domains, including algorithmic bias, defamation, intellectual property, and privacy. This Article specifically examines the implications of Al-assisted works, with a focus on text-to-image generators, such as Midjourney, that possess the ability to create detailed visual art from simple text prompts. The tools are not, however, limited to simple prompts. Users retain the ability to introduce greater complexity by specifying a host of variables that define the resulting image. Al-assisted art implicates significant legal rights and responsibilities. As to responsibilities: Can the image mislead, defame, or infringe on someone else's rights? As to rights: What are the free speech or intellectual property ownership implications? This Article interrogates the implications of these developments through the lens of copyright law, with a particular focus on the question of authorship. By analyzing how tools like Midjourney create visual art from textual prompts, this Article explores whether and to what extent these Al-generated outputs can be protected under copyright law. It describes current copyright doctrine through a novel framework, demonstrating that copyright law 's requirement for creativity pertains to an author's mental acts rather than the technical skill of fixation. The resulting requirement for conception-informed fixation is then applied to Al-assisted works, showing that an author can maintain sufficient creative control of an Al tool's output to meet the mark.