Artificial intelligence is already part of many music creation processes. It can be used to generate an entire song, create a voice, produce an instrumental part, or support specific stages of a work created primarily by humans.

‍

On July 10, 2026, organizations including IFPI, RIAA, A2IM, WIN, IMPALA, The Recording Academy, SAG-AFTRA, and the Human Artistry Campaign introduced a joint proposal to identify the use of generative artificial intelligence in recordings available on digital platforms.

‍

The system proposes two labels: AI-Generated and AI-Assisted.

Its goal is to provide clearer information about the process behind each song, without treating every use of this technology as the same.

‍

The AI-Generated label would apply to recordings in which artificial intelligence created all or most of the main creative elements.

‍

This could include, for example:

• an AI-generated lead vocal;
• a central instrumental performance generated by AI;
• a song created entirely from instructions or prompts.

‍

The AI-Assisted category, on the other hand, would apply to productions created primarily by humans that incorporate generative AI into certain creative elements.

‍

In these cases, the lead vocal and core instrumental parts are still performed by humans. AI is part of the process, but it does not replace the creative core of the recording.

‍

Using a tool to support one stage of production is not the same as releasing a song generated entirely through an AI platform. Both recordings may involve AI, but they do not reflect the same process, the same level of human involvement, or the same creative journey.

‍

Although the final result may be a visible icon displayed next to a song, the most significant change takes place before the release reaches a platform.

For a digital service to identify a recording as AI-generated or AI-assisted, that information must be included throughout the delivery process.

‍

This involves not only the people who create the music, but also labels, distributors, aggregators, digital services, and the organizations responsible for setting metadata standards.

‍

The organizations behind the initiative said they will work with these stakeholders to support its implementation across the industry. The system was designed as a voluntary framework that can evolve as technology and regulations continue to change.

‍

Transparency, therefore, will not depend solely on a platform’s technical ability to detect the use of AI. It will also require clear information from the moment a recording is uploaded and delivered.

For now, the labels apply only to the use of generative artificial intelligence within the sound recording itself.

‍

The proposal does not yet cover AI used in songwriting, lyrics, cover artwork, or music videos.

‍

There is also no confirmed universal rollout across all platforms. The labels are expected to become available soon, and their adoption will depend on collaboration between the different participants in the music ecosystem.

This means that several questions remain unresolved.

‍

One song may contain a small AI-generated element within an otherwise human-made production. Another may have been created using AI and later modified, performed, or produced by humans. Hybrid processes will become increasingly common and will not always fit neatly into two categories.

‍

Transparency does not determine whether a work has artistic value. Nor does it replace discussions around consent, intellectual property, attribution, or compensation. It does, however, allow audiences to better understand the origins of what they are listening to.

‍

In an environment where human-made music, AI-assisted music, and fully AI-generated content can coexist on the same platform, this information is becoming an increasingly relevant part of a release.

‍

Making that process visible could be one of the first steps toward building a clearer, more responsible music ecosystem that is better prepared for the changes ahead.