Essay
The Myth of the Isolated Work: Disputed AI Copyrights In Broader Entertainment Ecosystems
Debates about AI-generated music often focus on whether a particular output deserves copyright protection. Far less attention is paid to the larger structures that surround commercial works. A song is rarely just a song. It is branding, reputation, trademarks, contracts, and goodwill woven together into something far more complex than a single file.
In a previous essay, The Danger of Half-Measures, I argued that policies denying compensation while continuing to host and commercially benefit from AI-generated music create an uncomfortable contradiction. If a platform finds a work valuable enough to catalogue, distribute, and leverage as part of its service offering, questions naturally arise concerning the basis upon which compensation is withheld.
That discussion focused primarily on platform behaviour. The question worth considering now is broader.
Many contemporary debates surrounding artificial intelligence begin from a remarkably simple assumption: that a work can be examined as an isolated object detached from the commercial ecosystem to which it belongs.
Once reduced to that level, arguments frequently follow a familiar pattern. If an AI-assisted or AI-generated component lacks copyright protection, then the work itself is treated as possessing diminished status. Some go further, implying that uncertainty surrounding one aspect of a creation somehow weakens every surrounding claim associated with it.
The problem is that this assumption bears little resemblance to how creative industries actually function.
Modern commercial works rarely exist as standalone artifacts. They exist within networks of ownership, branding, reputation, contractual relationships, and consumer recognition. Those overlapping structures are not decorative features attached to a work after the fact. They are often among its most valuable components.
An audience member searching for a musician is not typically searching for an anonymous audio file. They are searching for a recognised artist. They may recognise a stage name, a logo, a visual identity, an album cover, a particular label, a website, a catalogue, or a reputation accumulated over years of activity. The commercial reality consists not merely of content but of identification.
That distinction matters because ownership has never been derived solely from copyright.
The public discussions surrounding artificial intelligence often compress an enormous collection of legal and commercial interests into a single question: is the output protected?
Even if that question were answered tomorrow, numerous others would remain.
Who owns the brand? Who owns the trademarks? Who owns the logos?
Who owns the consumer goodwill attached to a recognised identity? Who owns the reputation built through years of promotion, marketing, and public engagement? Who owns the contractual relationships connecting artists, distributors, labels, and audiences?
These interests do not cease to exist simply because one component of a larger work becomes legally controversial. Consider a hypothetical example involving a major entertainment company. Suppose an organisation creates a short film using significant AI assistance. One might reasonably debate the copyright status of various generated elements contained within that film. What does not logically follow is that the surrounding intellectual property suddenly becomes available for unrestricted commercial use.
If audiences recognise the characters, the logos, the branding, and the source of origin, those interests remain attached to the work. The uncertainty of one component does not transform every other component into public property.
Yet discussions surrounding AI frequently proceed as though ownership functions in precisely that manner. A guitar riff generated through software becomes the focal point, while everything surrounding it quietly disappears from view. We become so focused on the disputed brick that we forget we are standing in front of an entire building.
The future legal landscape will almost certainly involve overlapping layers of ownership that survive independently of one another. Some claims may succeed. Others may fail. Certain forms of copyright may evolve. New statutory frameworks may emerge. Existing precedents may be revised. Courts, legislatures, regulators, creators, distributors, and technology companies will continue negotiating where those boundaries ultimately belong.
A disputed brick does not negate ownership of the building.