Your work within this project will assess the extent to which current AI models incorporate specific copyrighted works and pioneer new methods for examining how traces of the training data appear within music embeddings and are regenerated in AI compositions. We are looking to quantify the influence of training examples on AI music compositions and use cutting-edge explainable AI techniques to analysing their encoding paths within the ‘brain’ of generative AI systems. During your contract, you will produce a prototype assessment model for artist remuneration based on diagnostic classifiers, which we will release as open-source software in tandem with a high-impact journal or conference paper.

Overall, we are striving to balance two competing societal objectives: fair remuneration of artists for using their work in training AI models that may potentially replace them at lower cost vs. ensuring that AI models are able to be trained on a sufficiently rich corpus of source material that artists will also benefit from revolutionary new tools that can support or even enhance their creativity.

Your tasks will be:

• design and conduct research under the guidance of the project leader, resulting in at least one peer-reviewed academic publication in international scientific conferences or journals;
• develop statistical models for assessing the relative contribution of training materials to generative music models and outputs;
• co-organise expert meetings or symposia involving industrial and legal partners in the field; and
• contribute collaboratively to the software infrastructure of the Amsterdam Music Lab.

• a PhD in artificial intelligence, machine learning, audio signal processing, music technology, or a related field (although candidates with a related master's degree and relevant industrial experience can also be considered);
• demonstrable experience working with recent AI models for music or audio generation;
• experience presenting scientific results at international conferences or publishing in leading international scientific journals;
• a strong cooperative attitude and willingness to engage in collaborative research;
• enthusiasm for communicating academic research to non-academic audiences;
• excellent social, communication, and organisational skills; and
• excellent command of written and spoken English.

• a 12-month researcher 4 (postdoc) contract starting 1 June 2025, with a probation period of 2 months;
• technical support from scientific programmers;
• excellent possibilities for further professional development and education;
• an enthusiastic and professional academic team;
• an inspiring academic and international work environment in the heart of Amsterdam; and
• the opportunity to collaborate with leading researchers and partners in the music industry.