Under existing copyright law, photographers own the rights to every
image they take instantly when the shutter clicks. That copyright survives if
a photographer alters an image using software like Adobe Photoshop. But
how might a photographer’s copyright be complicated by the use of
generative AI in altering their own source imagery? In March 2023, the U.S.
Copyright Office set out to answer this question by issuing its first formal
guidance on works containing AI-generated material. In it, the Copyright
Office attempted to delineate clear boundaries governing the copyrightability
of partially AI-generated works. But it fell woefully short. By relying too
heavily on existing doctrines and straining core copyright principles, the
Copyright Office’s March 2023 guidance extends insufficient protection to
partially AI-generated works and threatens to stifle creative expression just
as generative AI prompts a bold reimagining of authorship.
This Comment engages with the robust and ongoing discussion around
generative AI and copyright, scrutinizing in particular the U.S. Copyright
Office’s current approach to partially AI-generated pictorial, graphic, and
sculptural works. By leaning too heavily on existing derivative works and
compilation doctrines and utilizing outdated, poorly defined terminology, the
U.S. Copyright Office’s March 2023 guidance sows more confusion than
clarity and threatens to undermine core copyright principles. In lieu of
continued reliance upon its faulty guidance, this Comment suggests that the
Copyright Office broaden its longstanding human authorship requirement to
embrace hybrid human-AI authorship.
