The Algorithmic Persuasion event was an initiative of the Persuasive Communication division. It was organized on November 22 2019 at the University of Amsterdam by Brahim Zarouali, Sophie Boerman and Guda van Noort. There was great enthusiasm for the event, as approx. 25 participants attended the event. These attendees came from diverse universities from both Belgium and the Netherlands (including University of Antwerp, Ghent University, VU, UvA, Radboud University, and Erasmus University), diverse backgrounds (including Communication Science, Marketing and Law) and were at various levels in their career (ranging from PhD students to Full Professors).
The event kicked off with a keynote by prof. dr. Natali Helberger who is a university professor at the University of Amsterdam and holds a chair in Law and Digital Society with an emphasis on AI. She addressed algorithmic persuasion from a legal and ethical perspective and stressed the fine line between advertising and unfair manipulation. Moreover, she addressed venues for interdisciplinary research collaborations.
This keynote was followed by individual pitches by all participants, which showed a great and inspiring variety in research concerning algorithmic persuasion. The pitches covered different types of technologies (such as chatbots, news recommender systems, music recommenders) across different domains (commercial, health, and political communication, and law). These pitches allowed researchers to get to know each other.
The afternoon session started with a refreshing walk around campus, and a creative thinking session in which we discussed what the future of algorithmic persuasion looks like. A common and re-occurring idea was the two-sidedness of algorithmic persuasion. On the one hand, it creates a convenient, hyperconnected world. On the other, it decreases human autonomy and can have negative consequences to human well-being.
After the thinking session, two roundtable sessions were organized in which we discussed the societal implications of algorithmic persuasion, as well as the consequences and implications of algorithmic persuasion for theory, methods, and education.
With regard to societal challenges, a first important consideration was that data are not the same as algorithms: data are most of the time treated and collected in a biased way. Therefore, algorithms are by default biased as well. In terms of regulating algorithmic persuasion, it is important that regulations can be perceived as paternalistic and to consider who is a trusted regulating party. Literacy, empowerment, and transparency are key constructs.
Theoretical considerations that were discussed are that we might need to rethink or conventional theories, and that interdisciplinary research is not only interesting but also a necessity in examining this phenomenon. Methodological considerations were mainly related to data (how to treat trace and behavioral data), new modes of data collection (e.g., data donations), and research designs (involving experimental designs in a natural setting, while the setting is a moving target). Qualitative research would be needed to examine the production side of algorithms.
Educational implications discussed were that current programs should not be replaced but updated and extended. Not only explaining algorithms and algorithmic persuasion but also real-life experience with this phenomenon should be an integrative part of the teaching programs to learn. Reflective skills and attitudes with regard to societal challenges are important, for example reflections on the desirability and ability of technology.
All in all, it was an inspiring and productive event, in which we brought together a wide range of fellow researchers with shared, mutual interests in algorithmic persuasion. We engaged in open discussions about emerging questions and issues around algorithms in society, with a specific focus on methods, theory, education and their societal implications. The event provided participants valuable input for their own research lines, as well as an opportunity to establish fruitful future collaborations.