Ghent film composers demand AI safeguards for copyrighted music

Film composers and music-industry representatives meeting in Ghent called for stronger protection of copyrighted music as generative AI spreads through the sector, arguing that consent, payment and disclosure must govern the material used to train models. At the World Soundtrack Awards Film Music Days, held in Belgium from Oct. 8 to 10, the European Composer and Songwriter Alliance urged three principles: authorisation before copyrighted music is used for training, fair remuneration for creators, and platform disclosure of the catalogues and datasets behind their systems. The alliance, which represents more than 30,000 composers and songwriters through 57 member organisations, hosted a roundtable and received the event's Industry Award. Helienne Lindvall described authorisation, remuneration and transparency as essential to protecting those at the centre of music creation. She also warned musicians to read platforms' terms of service closely, saying uploads to services such as Suno can, depending on those terms, give a platform broad and perpetual rights to use and commercially exploit the material. The appeal came as the EU's AI Act, formally Regulation (EU) 2024/1689, established harmonised rules for artificial intelligence. The European Commission describes it as the world's first comprehensive AI legal framework. Its rules for general-purpose AI models took effect in August 2025, while the Commission lists August 2026 for the Act's transparency rules; the legislation includes transparency and copyright requirements for providers of those models. The debate at Ghent also centred on whether AI could supplant the human work from which film-music careers emerge. Simon Franglen, composer of “Avatar: Fire and Ash,” pointed to the hundreds of musicians who performed together for the score and questioned how AI could replicate the emotion created at a particular moment of recording. Panelists warned that generative tools and music-production factories could erode entry-level commissions, removing a foothold for emerging composers. Franglen cited Hans Zimmer's early opportunity composing for the 1980s British game show “Going for Gold,” while Darrell Alexander identified work as an additional composer for an established figure as a particularly valuable route into the profession. Adiescar Chase sought to hire further assistants for later seasons of Netflix's “Heartstopper.” Martin Phipps offered a more qualified view, saying creatives would continue to value human contact and that AI would not eliminate every job, though some directors might use it to compose music themselves. He added that prompting AI was itself a skill. The concerns coincide with broader uncertainty over copyright and model training. A 2026 European Audiovisual Observatory report says consent and potential harm to authors and rights holders are central issues when protected works are used without permission. It also finds that European and US approaches remain substantially different, that the EU has no AI-specific copyright legislation beyond its general copyright framework, and that the Court of Justice of the European Union has yet to establish a position on generative AI and copyright. A CISAC study reported by The Guardian projected that music-sector workers could lose almost a quarter of their income to AI within four years. It forecast generative-AI music accounting for about 20% of conventional streaming revenue and 60% of music-library revenue by 2028, while warning that creators face both unlicensed use of their works and fewer opportunities as AI-generated output competes with human-made music.