On Wednesday, January 26, 2022, Prof. Richard Wahlund, Bonnier Family Professor in Business Administration, hosted a workshop dedicated to AI-Driven Advertising at the Department of Marketing and Strategy at the Stockholm School of Economics (SSE). The workshop was part of an ongoing project on AI-Driven Contextual Communication funded by KAW, MMW, and WASP-HS, and engaged researchers in a plurality of scientific fields, including Computer Science, Statistics, and Economics. The purpose was to inform each other on ongoing research and identify shared points of interest. Adlede was represented by its co-founder and CTO, Johanna Björklund, who is also the primary investigator in the above-mentioned MMW project, by Mona Forsman, Head of R&D, and by Anton Eklund, industrial PhD student.
In the first session, Carl Piva, CEO for the national Internet foundation, talked about what his organisation can offer researchers, and Carl-Gustav Linden from the University of Helsinki followed with a lecture on metadata management at news agencies. Linden noted that there is a lack of consensus between publishers with respect to metadata taxonomies and annotation processes. He also remarked that manual annotations are not always a reliable gold standard in machine learning, as the journalists tend to prioritise the more creative parts of the workflow. The session was concluded by Sebastian Krakowski, SSE, on an experimental study of how humans work with AI in organisational decision making. Although still at an early stage, the study aims to provide a better understanding of how humans change their strategies when aided by AI-driven decision-support systems, and how their self-perceived understanding of the system's workings influences their willingness to trust their recommendations.
In the second session, Jonas Colliander at SSE talked about embedded e-commerce, where AI is used to insert shoppable panels into the publishers' editorial content, providing a native experience for the viewer. This creates new business models and revenue streams, but also raises concerns about the publishers' and advertising company's brand images. The next speaker was Anton Eklund, PhD student at Umeå University, who demonstrated a machine-learning system for clustering news articles based on topical similarity. This type of analysis is valuable for advertisers, as it gives them an understanding of what topics are being written about, and how they can position their offerings in relation to these. Eklund has set up an interactive demo here that is based on UK news data.
The final speaker was Anne Kaun from Södertörn University who talked about automated decision-making in the public sector from a citizen perspective. This theme covers digitalisation and resource economy, but also data privacy, trust, and fairness.
If you are interested to learn more or get involved in the MMW project, please reach out to Johanna Björklund at email@example.com.