Five PhD researchers will conduct research on A.I. at DPG Media.

An article by Lars Anderson and Gerbert Van Loenen

The timing couldn't be better. Three years ago, data manager Sven van Egmond, along with a (now departed) colleague from DPG, initiated the first contacts with universities to explore possibilities for collaboration between DPG and the world of science. This summer, at the peak of the generative AI hype, the collaboration is truly commencing as five PhD researchers are appointed within the company. Three positions have already been filled, while two remain open, including this position which can be applied for until September 1st.

With whom are we collaborating? Let’s start at the beginning: scientific research into AI in the Netherlands is spearheaded by ICAI, the Innovation Center for Artificial Intelligence. In their own words:

ICAI is a national initiative focused on joint talent and technology development between academic institutions, industry, and government organisations.

Think of it as a network of research laboratories, with each lab focusing on a specific research area. Over the next five years, companies and the government have allocated 87 million euros to fund seventeen labs. (Read an interview with Maarten de Rijke, head of ICAI, about what they do.)

One of those labs is the Responsible Media Lab (REM Lab), in which DPG Media will collaborate with the Jheronimus Academy of Data Science (JADS) and the University of Amsterdam (UvA). DPG is investing one million euros in this lab, another part comes from government subsidies.

The goal of this lab is to explore the possibilities of AI-driven solutions for media. This includes professionalisation, new business models, and ways to engage with the audience. (Source: REM Lab.)

The five PhD positions focus on the following topics:

  • Legal aspects of using (generative) AI in media. Key question: How do we translate the legal requirements of transparency and explainability for news recommendation systems?

  • Explaining recommendations. Key question: How can recommendations from algorithmic results, journalistic choices, and company rules be explained to users in an understandable manner?

  • Synergies between platforms. Key question: How do we identify new metrics to measure network effects within and between platforms, aiding in predicting new synergies between platforms?

  • Network perspective on user content and behaviour. Key question: What role does media content play in shaping the preferences and behaviour of the audience?

  • Content generation. Key question: Can neural NLG (natural language generation) produce qualitative, reliable, configurable news? Reliability requires clear, factual, relevant, and ethical news. Quality pertains to reader satisfaction and minimal editorial time. Configurability allows editors to personalise the style and structure of the content.

The five PhD researchers will conduct four to five years of research within this lab, and will be partially (at least one day a week) based in Den Bosch, sharing their knowledge within the company. DPG provides the context, data, and infrastructure necessary for the research projects. The researchers can work with a panel of approximately fifty thousand DPG users.