Ibán García del Blanco, secretary‑general of the Association of European Performers' Organisations (AEPO‑ARTIS), told UnionPress that the EU's artificial‑intelligence legislation fails to protect the rights of musicians, actors and dancers whose performances are harvested to train generative systems.

He argues that, unlike authors who own the underlying composition or script, performers rely on a fragile set of "neighbouring rights" that safeguard the recording of their contribution. Those rights, he says, are being eroded as record labels, distributors and streaming platforms sign licences with AI firms without consulting the artists whose voices or movements are being digitised.

Neighbouring rights explained

Neighbouring rights are a European‑wide legal construct that give performers a personal stake in any sound or visual recording of their work. While the copyright in a symphony belongs to the composer, the oboe player who brings the melody to life retains a right to prevent the unauthorised exploitation of that specific performance. In practice, the chain runs from the orchestra to a record label, then to a distributor, and finally to any third‑party that wishes to use the recording, including AI developers.

García del Blanco points out that the current system allows a distributor to grant a licence to an AI company for training purposes, even if the performer has not been consulted. "The performer's reputation can be damaged if the AI reproduces a flawed or out‑of‑context version of their work," he says, noting that neighbouring rights were designed to protect against exactly that scenario.

AI as a new threat to the performing arts

According to the AEPO‑ARTIS chief, AI is already reshaping the economics of the cultural sector. "We lack a real impact assessment of what is happening now and, more importantly, what will happen in the future," he warns. The concern is not merely theoretical; AI models are being fed recordings of live concerts, theatre productions and dance performances to generate new pieces that can be sold or streamed without any remuneration to the original artists.

For many performers, especially those who are not headline stars, the loss of income from such unlicensed uses could be significant. "The position of performers is already less protected than that of authors," García del Blanco explains. "If you compare music and audiovisual performers, the protection and guarantees are not the same." He adds that the disparity is even sharper for freelance musicians and dancers who depend on gig work and have little bargaining power with large production companies.

EU AI rules miss the mark

García del Blanco was part of the team that negotiated the EU's AI Act, which aims to set standards for high‑risk AI systems. He says the legislation focuses on transparency and safety for consumers but overlooks the specific needs of cultural creators. "Policymakers treated AI as a generic technology problem, not as a tool that can appropriate creative labour," he argues.

The Act does require providers of high‑risk AI to conduct conformity assessments and to respect fundamental rights, but it does not obligate them to obtain explicit consent from performers whose recordings are used for training. As a result, AI firms can legally scrape large music libraries or video archives that are already under licence for other purposes, and then repurpose the data without additional remuneration.

Union representatives across Europe have begun to lobby for an amendment that would extend the scope of the AI Act to cover neighbouring rights explicitly. They propose a mandatory "fair‑share" clause that would require AI developers to negotiate licences with the performers' collecting societies before using any recorded material.

What unions and industry bodies say

Several national performers' unions have echoed García del Blanco's concerns. The German Musicians' Union (Ver.di) recently issued a statement calling for "robust safeguards that prevent the exploitation of live recordings for AI training without fair compensation." In France, the Syndicat National des Artistes (SNA) has launched a petition demanding that the French Ministry of Culture enforce stricter consent rules for AI developers.

On the other side, some industry groups argue that overly restrictive rules could stifle innovation. The European Recording Industry Association (ERA) claims that a "balanced approach" is needed to protect both creators and the development of AI‑driven tools that can help artists with composition, rehearsal and marketing.

García del Blanco acknowledges the potential benefits of AI but stresses that they must not come at the expense of human performers. "I am not in favour of breaking computers, but we must ensure that human creativity and human beings will continue to have a place in our lives and are not replaced by tools that, at the end of the day, imitate more than they create something new," he says.

Possible paths forward

One proposal gaining traction is the creation of a pan‑European digital rights register that would list every performance recording and its associated neighbouring rights holder. Such a database could streamline the licensing process for AI developers while guaranteeing that performers receive a share of any downstream revenue.

Another avenue is the extension of the EU's existing collective‑management framework to cover AI training data. This would allow societies like GEMA in Germany or SACEM in France to negotiate blanket licences on behalf of their members, ensuring that a portion of the fees collected from AI firms is redistributed to the artists.

In the meantime, AEPO‑ARTIS is urging performers to be vigilant about the contracts they sign with record labels and distributors. "Artists need to ask whether the label's licence includes the right to sell the recording to AI companies," García del Blanco advises. He also recommends that performers seek legal advice before entering agreements that could waive their neighbouring rights.

As the EU prepares to finalize the AI Act later this year, the debate over how to protect the cultural sector is likely to intensify. For many musicians, actors and dancers, the outcome will determine whether they can continue to earn a living from their craft or watch it become a data point for ever‑more sophisticated algorithms.

"We are at a crossroads," says García del Blanco. "Either we adapt the rules to give performers a fair share in the AI economy, or we risk seeing a generation of artists pushed out of the market altogether."