Algorithmic imaginaries. Visions and Values in the shaping of search engines

Search engines like Google are developed in the US-American context, but are used around the globe. Their business models are based on user-targeted advertising. They collect user data, turn it into user profiles, and sell them to advertising clients. Since the NSA affair practices of user profiling are critically discussed; especially in European contexts with diverse data protection laws, historically shaped notions of privacy, and very different tax systems. The ongoing reform of the EU data protection legislation is an important arena where tensions between global search engines and European policy visions and values can be observed. Besides, European search engines emerge that aim to provide users with alternative styles of search. Some are explicitly developed as a European competitor to US-based search engines (Quaero or Independent Web Index). Others are developed in Europe, but draw on other value-systems to distinguish themselves from big search engines, such as respecting users privacy (e.g. Ixquick), protecting the environment (e.g. Ecosia), or creating a non- commercial search engine owned by the public (e.g. YaCy). This poses important questions: What motivations, value-systems, and visions guide the development of European search engines? How are these imaginations translated into sociotechnical design practices? What power struggles, negotiations, and compromises may be observed? How do place and cultural context matter in the design process? Researchers in Science and Technology Studies (STS) investigated the politics of search engines, the relevance of algorithms, and internet governance. What is missing is an in-depth analysis of the shaping of search engines in specific cultural contexts and the role shared value systems and visions play in it. Rooted in the discipline of STS the suggested habilitation project will fill in this gap by investigating design practices of European search projects using a case-study approach (interviews, workshops, observations). Results from this analysis will be compared to and cross-analyzed with results from my past research on capitalist ideologies driving global search engines like Google and my present research on visions and values guiding European search engine governance. This overall analysis will result in a typology of algorithmic imaginaries, which describes visions and values in the development and governance of search engines in global, European, and local contexts. It will show how search technologies and society co-emerge in specific economic, political, and cultural settings. The primary focus on European contexts is a particular strength of the project since tensions between global search engines and European governance structures and search projects are growing, but have not been systematically studied yet, both in the field of STS and internet research.

The project investigated three alternative search engines from Europe: The privacy-friendly search engine Startpage, the peer-to-peer search engine YaCy, and the Open Web Index initiative. The analysis focused on visions and values shaping these search engines, how these get translated into search technology, and how the European context matters in these practices. The results show that the search engine projects are driven by values such as privacy, openness, de-centrality and independence. These values are not fixed or rigid, but rather fluid, context-dependent, and changing over time tightly intertwined with the development of the technologies. This flexiblity enables the projects to develop a certain value pragmatics needed to grow and become more sustainable. Moreover, "European values", and broader notions of Europe as "unified or pluralistic", were imagined and co-produced along with the technologies. Some of them, such as privacy and digital sovereignty, were anchored in larger European narratives to situate and promote the projects. Furthermore, less prominent visions of Europe were shaped in the context of alternative search engines, which pointed towards challenges in the particular Europen context. The frame of "bureaucratic Europe" most importantly, which search engine developers related to cumbersome funding structures and a reluctant start-up mentality. Moreover, alternative visions of a pluralistic Europe were articulated in regard to techonlogical diversity and de-centrality. An open web index, for example, could create a whole range of different search engines, ranking algorithms, and applications. This would better correspond to multicultural, divers, and federal European contexts than big tech companies like Google that primarily count on monopolism and commercialization. To conclude, the project finally suggested three interventions that may help pave the way towards pluarlistic Europe based on technological diversity and decentrality. Long-term funding and slow scalability: Big infrastructure projects like an open web index as a backbone to search engine diversity require bold, long-term, and sustainable funding structures to develop digital sovereignty. Continuous auditing and advice: The making and governing of digital technologies requires continuous and interdisciplinary advice guaranteed by professional, independent processes and authorities, as well as adequate resources; instead of ex-post, snapshot auditing. Opening up data? The development of alternative technologies and infrastructures requires data to train algorithms and machine learning models. How to open up commercial data, to share public data, and to create collective data pools that go beyond individual responsibility and ownership of data are central questions that need to be tackled in the future. The results of this project will be combined with results from two previous search engine projects to write the habilitation "Algorithmic Imaginaries". This habilitation will show how search technology and society co-emerge in specific economic, political, and cultural environments; with a particular focus on the European context.