In this research project, we want to explore the possible advantages and disadvantages of future virtual companions, delivering recommendations and suggestions for legislation that are well anchored in a stakeholder community both nationally and internationally. Our focus is not on developing new technology, but instead investigating applied information technology, that is, how the technology fits in and interacts with society and individuals in their daily lives.
Virtual companions, as we define it, are conversational agents that primarily have the purpose of being a social and continuous companion to a human user. These agents can be based on large language models, or other AI or chatbot technologies. The user primarily interacts with the AI through text messages, but it can also have a visual representation in the form of an avatar, and a voice generated by speech synthesizers. Their function can take many different forms, ranging from emulating friendship to relationships that are, for example, therapeutic or romantic in nature.
There are some existing examples of companion chatbots, such as Replika, but both from a technology and business perspective they are still in their early phases of development. We believe that the development and usage of virtual companions will reach a hype peak in just a few years, and then a maturity plateau in about 15 years.
Already, systems such as ChatGPT enable a dialog where the user can ask questions and it responds in a human-like way. We predict that there will be an increase in both the usage and quality of these chatbots in the future, but it is hard to say exactly to what extent. We imagine a future where virtual companions are more advanced and more human-like, allowing people to talk to them through their smartphones, computers, and other devices, and not only to ask questions and receive help, but also to use them instead of, or as complements to, human relationships. Obviously, this remains to be seen, but if this leads to a large-scale adoption of virtual companions in society, then we could anticipate profound effects on people’s lives, behavior, social relations, and ultimately their wellbeing.
What is unknown at this point is how these chatbots affect individuals, especially adolescents and young adults. These effects could be similar in scale to how social media has affected society and individuals, and we want to investigate the phenomenon before it occurs, preparing society for the possibility of disruptive change. We want to focus on the mental health and wellbeing of adolescents and young adults, primarily based on how social media seems to have had negative effects on especially that group of users.
Our current research question is: What advantages and disadvantages can we expect from future development of virtual companions, and how should we as a society harness the advantages and avoid the disadvantages in the coming two decades?
How do we explore something that hasn't happened yet? First, we will explore people’s attitudes and conceptions about virtual companions, using ethnography, primarily in the form of interviews. Then, we will use design fiction as a method to explore potential future developments. Design fiction means developing expanded scenarios with a high level of details and deep characterizations of the user’s behavior and situation. This can be done in literary form, as short films, or as stage plays and role-playing games. We will curate and publish design fiction from invited storytellers, for example in the form of anthologies or small film festivals. Third and finally, we will use participatory action research, meaning that we involve stakeholders in workshops where the design fiction scenarios will inform and inspire the work of formulating recommendations and suggesting legislation.
In order to create a major impact in society, we plan to involve stakeholders with extensive authority over the development of society, such as experts on mental health, politicians, company leaders, and top managers at government agencies.
The expected deliverables from the research project (beyond the final PhD thesis) are A) insights among the stakeholders that have participated, B) written recommendations directed at both people developing the technology and as well as people using the technology (such as psychiatrists working with adolescents and young adults), and C) suggestions for future legislation.
Contact
This research project is conducted by PhD student Negin Hashmati, and supervisors Thommy Eriksson, Mohammad Obaid, and Therése Julia Iréne Skoog.
If you are interested in our project, please feel free to contact Negin hashmati@chalmers.se or Thommy thommy@chalmers.se
