2nd December 2025When two become one: the impending challenges of auditing automated media mash-ups and remixes

Digital Service Providers are experimenting with AI-driven re-composition, which automates the blending and transformation of existing recordings into composite new works. This algorithmic innovation in music streaming appears to mark a shift from simple recommendation to re-composition. As platforms begin to be able to automate this process, a complex interplay emerges between metadata structures, royalty pipelines and a host of impacts on existing contractual clauses which create compliance challenges for all parties.

Algorithmic curations of music mash-ups and remixes

In our last article “Hard Clauses, Soft Constraints” we noted the evolving challenge of auditing algorithmic outcomes and marketing and positioning clauses associated with how DSP algorithms curate content. However, some patents suggest DSP algorithms are moving beyond curation into re-composition with the move towards automated mash-ups and remixes, which blend tracks or audio segments to create hybrid works.

It’s easy to see how such innovations may be popular with users, and drive engagement through the near infinite potential for novelty. If this does increase engagement there’s great potential for this to translate to greater income for DSPs and licensors alike. However, as the discrete relationship between an asset and a play severs, this creates additional risks for license holders and auditors alike to be aware of, in terms of metadata and royalty pipelines. What’s more, these hybrid works create challenges regarding interaction with other contractual clauses, which may impact both income allocation and the visibility of content.

From discrete to multi-asset metadata

When a user logs onto their platform of choice and selects a playlist or specific track or specific video, that interaction will be logged with a particular event id (e.g. stream ID), along with the respective content ID. This interaction between the user and the content is a discrete event; should the user skip and select another piece of content, then that event ends and the duration of the users’ interaction with this asset is logged, with the play of the subsequent content comprising another discrete interaction (with appropriate metadata).

Most DSPs distribute catalogues provided by licensors, ingesting the associated catalogue metadata. Whilst the DSPs platforms impact visibility of the provided content, most of the assets are simply distributed as provided and as such, an event ID translates to a single asset ID (and also to a single ISRC in the case of a music recording). But some DSPs do distribute content that comprises of multiple distinct assets (e.g. a video uploaded by a user with multiple songs), and the metadata and relational database queries in the royalty pipeline should reflect this distinction. This is typical of DSPs who must factor in user-generated content (“UGC”).

Accordingly, algorithmic mash-ups and remixes, being multi-asset content, will likely pose an iterative challenge to platforms that already handle existing multi-asset content ( with multiple songs). However, they present a potentially larger data architecture challenge for platforms whose back-end processes were created around discrete content units. Both an increase in complexity of the DSP’s royalty pipeline, and an increase in the complexity of revenue/royalty calculation inputs increase the challenge of compliance for DSPs launching such modes of consumption. This also increases the value an audit of outcomes associated with such royalty pipelines changes.

Interactions with contractual clauses

In addition to the technical challenges in implementing such changes, an additional challenge will be how such a change interacts with existing contractual clauses. For licensors, the following questions will be particularly pertinent in assessing compliance risk:

• How are tracks for mash-ups and remixes selected? Are they licensor agnostic?
• How will they interact with existing marketing clauses? If we have clauses that promote our tracks with respects to playlists and recommendations, potentially in exchange for a lower effective royalty rate, will such clauses be considered when content is selected for a mash-up?
• If such mash-ups are considered part of the respected service for which an aggregate outcome is contractually defined, and if there are lower effective royalty rates applicable to this new content type, could an over-representation of such views comply with aggregate outcome clauses and yet still have a negative overall impact on income?
• Similarly, if there is a positioning clause (specifying location of placement of content), could such a clause be complied with via a mash-up and remix whilst only generating an apportioned stream (e.g. half a stream)?

The exact compliance challenges, and testing to be undertaken to verify compliance, will depend on the answers to the above questions.

Auditor and licensor takeaway

A substantial change in how content is offered on a DSP platform may result in significant changes to the metadata structure and royalty pipelines. Additionally, the interaction between existing clauses and definitions (such as marketing and positioning clauses, or even derivative components of market share calculations) will likely lead to additional complexity.

This additional complexity naturally creates a greater compliance risk for DSPs. Licensors, DSPs and auditors should be aware of the increased risk, and work together to mitigate and address those that impact contractual compliance.