Algorithms and digital media

Search engines, social networks and online media are a part of everyday life for many people, especially those looking for information online. These digital media are usually controlled by algorithms, the exact function of which is often not known to the user. For example, the result of a Google search or a Facebook newsfeed algorithm varies according to previous searches and individual user data. Thus, there is a serious possibility that personalized, algorithmic mechanisms can influence and even manipulate public opinion. Even traditional media companies have been relying more heavily on algorithms. The aim of this study is to determine the effect of algorithms of the formation of public opinion. An online survey as well as multiple interviews with first-time voters will complement the study.
© Weissblick -
Algorithms in digital media and their impact on public opinion
2017 - 2018
German Bundestag/Office of Technology Assessment at the German Bundestag (TAB)


Online media are an important source of political information and a basis for the formation of opinions. An increasing number of people also use them to familiarize themselves with relevant social issues. Search engines are an essential component of this process as they, along with websites and social networks, form the mainstay of the population’s daily media usage.

The operating principle of digital media is determined by algorithms that can decide which websites are deemed relevant for a given query, as well as which order they are presented in. They form the basis for media usage data, choosing which selection of news reports and advertisements a user is shown at any given time. Through this, they can influence the opinions of their users.

Not only social media is effected by this, as algorithms are even growing in prominence within more traditional forms of news. Journalists, as well as other individuals that affect public opinion, are supported by algorithms while performing tasks such as researching, selecting and evaluating information, as well as while writing and distributing articles. While automated messages are still relegated to areas in which standardized data is prevalent, such as scoreboards and corporate reports, the development and rapid expansion of self-learning algorithms may soon result in a spread to other areas.

Algorithms are also used to explore the preferences of potential recipients. For example, the Washington Post and the US news portal Upworthy show multiple versions of the same article to a group of test subjects for review before publishing it. Specialist software is then used to determine the optimal combination of title, images and text modules for the target audience.

It can be assumed that algorithms are becoming increasingly important when it comes to forming public opinion, however while they allow the consumer to access a much wider spectrum of information and (political) positions compared to older forms of media, there is the risk of them influencing and manipulating opinions through subtle algorithmic mechanisms, whether intentionally or otherwise.

In recent years, the discussion over the complex interaction of algorithms and digital media has moved from specialist circles to the public eye, mostly thanks to mass media. This results in the discussion over important issues, such as: Which content do Facebook users see in their news feed? Which results are prioritized in a Google search? How do media providers use the new possibilities for the creation and prioritization of news coverage? How much does algorithmically supported media coverage influence public opinion? Can fake news, hate speech and filter bubbles impact democratic decisions?

In 2016, an interdisciplinary team of scientists discovered that there were far too few analyses for the potential impact of our usage of algorithms on the social, cultural and political sectors. Civil society organizations demand better transparency and possibly increased control regarding algorithms.

Goals and approach
In the TAB project, the complex subject was approached from two directions: The dynamic, algorithmically influenced developments were discussed in relation to traditional as well as social media, with a focal point of discussion being their effect on shaping opinions. The starting point consists of literary analyses of the technical basis and the usage of algorithms in digital media, for scientific discussion and potential political means of action. The public discussion will be assessed through a variety of contributions from print media and radio. Interviews with experts from scientific and practical fields will support this approach.

Selected aspects should be discussed in depth in the context of studies and surveys of the stakeholder panel TA (technology assessment). For this purpose, discussions will be held in focus groups with theses generated for questions concerning the potential benefits, limitations, opportunities and risks users associate with personalized messages. An online survey will verify the results of the focus groups.

With this TA-project, the TAB follows on the TA-preliminary study “Social Bots”, which concerned itself with the effect on social bots on public opinion, as well as other consequences of their use. Social bots are computer programs whose purpose is to generate automated content on social media to influence public discussion.


  • 2018

  • Kluge, Jakob; Oertel, Britta; Evers-Wölk, Michaela (2018): Wie beurteilen junge Menschen personalisierte Onlinemedien?. Ergebnisse einer Repräsentativbefragung. Büro für Technikfolgen-Abschätzung beim Deutschen Bundestag. Berlin. (TAB-Sensor, 1).

    Wie beurteilen junge Menschen personalisierte Onlinemedien?. Ergebnisse einer Repräsentativbefragung

    Export record: Citavi Endnote RIS ISI BibTeX WordXML

Share contents: