Our co-director Natali Helberger was recently part of a panel organized by The Data Tank, a nonprofit “think and do tank” based in Brussels focused on the responsible use of data for the common good. The conversation revolved around how EU policy can support public interest media organizations in dealing with the entrance of generative AI on the market and the normative power that large providers of such systems have over sector dynamics. The Data Tank is conducting a line of research on this topic, and their recently published advisory report was discussed by Natali and fellow panelists Saorla McCabe (UNESCO), Maria Luisa Stasi (Council of Europe & Article 19), Paul Keller (Open Future), Tamian Derivry, (European Commission); Mira Milosevic (Global Forum for Media Development), and Wout Van Wijk (News Media Europe).
The current changes of GenAI bring deep and structural power asymmetries. As the panel noted, these require holistic solutions that strengthen collective bargaining power, and approaches that don’t leave behind local, niche, or investigative outlets that provide important input but may get overlooked.
Some key points made by Natali were that the availability of choice between technological solutions, including public interest alternatives, is a prerequisite to establishing fair negotiation power. Existing risk-based regulation may reduce dependencies, but it will not make the technology better or more democratic. She highlighted that an important part of the audience’s freedom of expression is the right to benefit from innovation, and states have a positive obligation to enable that. Natali also indicates that the complexity of the topic can have a disheartening effect on individual organizations, and suggests to collectively formulate concrete action points as next steps to help set change into motion.
To this point, the panel converged towards the end of the session on the idea that technological innovation needs to be redefined as a collaborative, democratic project that works in the public interest and across national borders. In this way, generative AI can be developed locally to do more than just mitigate the risk of dependence, but also to more closely serve the needs of its users and respect their media freedom.
The Data Tank’s research is based on a multi-stage process involving a review of literature, ecosystem mapping, and consultations with practitioners and stakeholders at large; click below for an overview of their research as well as their final report.
More info:
