Using AI for social benefit – WeNet and community focused social media

Using AI for social benefit – WeNet and community focused social media

The internet has proven to be a successful mechanism at finding social and cultural commonalities across space and time, but what would an internet based on highlighting diversity look like and what social dynamics would this create? Reporting findings from the WeNet study, George Gaskell, Amalia de Gotzen, Amarasanna Ganbold, Sally Stares and Katharina Durmeier discuss how the project aimed to create a more diverse and community focused form of digital connection.

As mainstream social media becomes fragmented, focused on ‘pay to play’ models and subject to disinformation campaigns, their original promise of a utopian new public sphere can feel far from reality. Was this inevitable, or are there ways in which digital media can be more diverse and community focused? Funded by the European Commission, the WeNet study keeps the flag flying for a more sociable form of digital connection.

The study started from a simple idea.  Confronted by a challenge or a problem, people might turn to their social circle, family, friends, and colleagues in the hope that someone will have the know-how to help. Self-evidently, the larger and more diverse the social circle, the more likely it is to include someone who is competent to address the issue. Furthermore, if the issue has multiple valid answers, for example “what’s the best book on existentialism?” a large and diverse social circle may offer a range of suggestions including surprises – ”forget existentialism, read Montaigne”.  Diversity may open the door to serendipity.

A central aim of the WeNet study was to develop and release as open source a general, application-agnostic platform designed to collect data that capitalises on diversity. And in parallel to create a data infrastructure comprising person-centric and behavioural data for use in empirical studies aimed at supporting new approaches to better understand and serve diverse communities.

The WeNet’s chat app, developed by the consortium harnesses diversity-aware artificial intelligence to exploit the distributed competences and interests within a community, empowering social interactions beyond any one individual’s social circle. In this way the chat app extends a person’s social circle to a community in a virtual world to the benefit of all. The chat app was developed and tested over three iterations each spanning three weeks with students in the social sciences at the London School of Economics and Political Science, industrial design at Aalborg University, and computer science at the National University of Mongolia.

The WeNet’s chat app, developed by the consortium harnesses diversity-aware artificial intelligence to exploit the distributed competences and interests within a community, empowering social interactions beyond any one individual’s social circle.

Participants ask a question by triggering the chat application’s question command. They are then asked to specify the domain of their inquiry (music, local things, arts, career, academic studies, cultural interests etc.). The question asker can signal that the question is sensitive (the chat app lets the answer givers know). The algorithm then selects five members of the community whose profile matches the question content, to whom the question is sent.

Students’ personal profiles were elicited in an online survey with questions based on social practice theory’s conceptual triumvirate: materiality – tangible assets; competences – skills and knowledge and meaning – values and beliefs.  Measurement of these personal characteristics defines each participant as a point in a multi-dimensional space that maps the diversity of the community, allowing for the algorithm to select appropriate respondents for a particular question.

The WeNet study had a strong commitment to research ethics and the ethics of social media platforms. Messages reminded participants of the importance of open, tolerant, and appreciative communication in interactions in the community.  When a participant had a question involving a sensitive issue, they could opt to be anonymous. Privacy and data protection went beyond legal requirements of the EU’s General Data Protection Regulations with the study adopting a minimalist approach to collecting sensitive characteristics such as gender, sexuality, ethnicity, or disability.

A coding of the questions submitted in the second two pilot studies (in months 35 and 46 of the WeNet project) shows a wide range of topics including requests for recommendations and suggestions about recipes, books, places to visit, chatting and small talk, and advice on academic issues (see Table.1).

Table.1: Types of questions posed in month 35 (M35) and month 46 (M46) pilot studies. Figures are percentages within each pilot study, within each university.

Exit surveys and focus groups with the student participants were conducted.  Asked for three words to summarise their experience, ‘fun’, ‘helpful’, ‘informative’ and ‘innovative’ were the most common descriptors. The surveys showed that giving advice tended to be more satisfying than receiving it. Badges awarded automatically in the app for the frequency of providing answers were much liked and encouraged greater participation, and the range and diversity of answers was appreciated.

The exit studies provide robust empirical evidence that the chat app was attractive to multiple student user types, for multiple uses and in different cultures. Table.2 provides information on numbers of participants in the pilots, numbers of questions posed in each study, and rates of overall satisfaction reported by participants. It indicates that rates of satisfaction generally increased over the course of the three pilot studies.

1698841391 292 Using AI for social benefit – WeNet and community focused | lifefromnature

Table.2: Numbers of participants, questions posed and rates of satisfaction with the chat app, for each pilot study (at months 26, 35 and 45) in each of three universities

Participants said that the distinguishing features of the chat app were the absence of advertisements and photographs, the opportunity to be anonymous, that it was limited to a trusted community, and the pleasure of helping and advising others in the student community. A final question in the exit survey asked the participants whether “Using the chat app in this university would benefit students”. The percentage agreement was 94 at the LSE, 78 at Aalborg and 83 in Mongolia; an overwhelming vote of confidence from users.

In addition to the three university pilot studies the chat app has been adopted successfully by a Spanish NGO that coordinates a network of more than 1,800 cyber volunteers helping people with computer problems. Another user is a Greek social cooperative enterprise which aimed at matching social entrepreneurs, who would like to start or join a social cooperative and are looking for either the right team or start up mentoring. All in all, the chat app is an attractive social media platform for different communities and in different cultures.


The WeNet project was funded by the European Commission. General information about the project can be found on the project webpage:

The content generated on this blog is for information purposes only. This Article gives the views and opinions of the authors and does not reflect the views and opinions of the Impact of Social Science blog (the blog), nor of the London School of Economics and Political Science. Please review our comments policy if you have any concerns on posting a comment below.

Image Credit: Adapted from QPT on Shutterstock.

Print Friendly, PDF & Email