Skip to content

What ValueMonitor can do

Differences between social realms in how digital values are addressed

We analyzed how often 12 different values that are relevant for digital technologies were mentioned in four different datasets. The first dataset (NEWS) was composed of a large number of news articles (#562.295), taken from 26 different newspapers (including Reuters, CNBC, The New York Times). This dataset is used to evaluate which values are associated with digital technologies in public debates. The second dataset (ETHICS) consists of 8.565 scientific articles downloaded from ethics-related journals. We have filtered both the NEWS and ETHICS datasets on articles discussing relevant digital technologies. The third dataset (TECH) contains 674.656 scientific articles downloaded from Scopus, selected using keywords related to relevant digital technologies, and excluding ethics-related journals. The fourth dataset (LEGAL) has been built using regulatory documents related to relevant digital technologies found on the EU Legislative Observatory website.

The figures show which values are most frequently mentioned in our datasets. In the NEWS dataset the most discussed values are democracy, cybersecurity, and justice & fairness. In the ETHICS dataset, the values of democracy and transparency are most often mentioned. The values most frequently addressed in the TECH dataset are cybersecurity and reliability. Finally, in the LEGAL dataset, the most prominent values are cybersecurity, transparency and justice & fairness.

Notably, some values are very prominent in some datasets while significantly less frequent in others. This is the case for the value democracy, which is an important value in the NEWS and ETHICS dataset, while being one of the least frequently mentioned values in TECH and LEGAL. Transparency seems to be a frequently mentioned value in ETHICS and LEGAL, while playing a smaller role in NEWS and TECH. Finally, well-being is in all the 4 datasets among the least frequently mentioned values.

It should be noted that frequencies with which values are mentioned cannot be directly compared between datasets because one would not expect values to be mentioned as often in technoscientific articles as, for example, in the news. However, more telling is the relative importance between values in different datasets. The fact that democracy is the most frequent value in NEWS and one of the least frequent in TECH clearly seems relevant. A possible explanation for this difference might be that, unlike other values like privacy and cybersecurity, democracy is a value that mainly needs to be addressed through regulation and legislation rather than through technological innovation. For example, by manually going through documents on AI and democracy in the NEWS dataset, we encounter discussions of potential threats caused by AI on democratic processes and its effects on people‚Äôs political opinions.  It might be that such threats are hard to address through technological solutions, and primarily require regulation. However, a manual investigation of articles on artificial intelligence and democracy in the TECH dataset shows that articles tend to discuss supervised learning algorithms and natural language processing as potentially helpful for classifying political opinions and detecting fake news. This suggests that there are also potential technological innovations and research priorities to better address the value of democracy. All in all, our analysis might point out that there is a gap in the technoscientific literature in how often the value of democracy is considered in conjunction with digital technologies.

The analysis shown is part of the report Ethical and societal challenges of the approaching technological storm written for the Panel for the Future of Science and Technology of the European Parliament.