Simsimi Chatbot

Unique Characteristics of Human-Chatbot Conversations and Their Potential For Mental Health Support. 

Accessibility and availability of psychological support is crucial for the mental well-being of the global population. The recent popularity of social artificial agents, such as chatbots, reveals the rising potential for using them for mental healthcare. In this work, we investigate how differently users express emotions online when interacting with a commercially available chatbot (Simsimi) and posting on a public social media platform (Twitter). Our results suggest that private conversations with a chatbot tend to be more emotionally charged and focus on personal issues more frequently than messages posted on Twitter. 

 This is an ongoing project in collaboration with Simsimi corporation. 

Sleep Culture

While sleep positively impacts well-being, health, and productivity, the effects of societal factors on sleep remain underexplored. Here we analyze the sleep of individuals across the globe from wearable devices and attempt to understand how culture affects our sleeping habits. Defining the interplay between social norms and sleep may allow the development of targeted interventions and health policies.

Park, S., Zhunis, A., Constantinides, M. et al. Social dimensions impact individual sleep quantity and quality. Sci Rep 13, 9681 (2023).


Presentation at TheWebConf 2022

Emotion Bubbles

The COVID-19 pandemic has been the single most important global agenda in the past two years. In addition to its health and economic impacts, it has affected people’s psychological states, including a rise in depression and domestic violence. 

We traced how the overall emotional states of individual Twitter users changed before and after the pandemic. Our data, including more than 9 million tweets posted by 9,493 users, suggest that the threat posed by the virus did not upset the emotional equilibrium of social media. In early 2020, COVID-related tweets skyrocketed in number and were filled with negative emotions; however, this emotional outburst was short-lived. We found that users who had expressed positive emotions in the pre-COVID period remained positive after the initial outbreak, while the opposite was true for those who regularly expressed negative emotions. Individuals achieved such emotional consistency by selectively focusing on emotion-reinforcing topics. The implications are discussed in light of an emotionally motivated confirmation bias, which we conceptualize as emotion bubbles that demonstrate the public’s resilience to a global health risk.

[Paper] [Presentation]

Covid-19 Simulation

The COVID-19 pandemic left its unique mark on the 21st century as one of the most significant disasters in history, triggering governments all over the world to respond with a wide range of interventions. However, these restrictions come with a substantial price tag. It is crucial for governments to form anti-virus strategies that balance the trade-off between protecting public health and minimizing the economic cost. This work proposes a probabilistic programming method to quantify the efficiency of major non-pharmaceutical interventions. 

We present a generative simulation model that accounts for the economic and human capital cost of adopting such strategies, and provide an end-to-end pipeline to simulate the virus spread and the incurred loss of various policy combinations. By investigating the national response in 10 countries covering four continents, we found that social distancing coupled with contact tracing is the most successful policy, reducing the virus transmission rate by 96% along with a 98% reduction in economic and human capital loss. Together with experimental results, we open-sourced a framework to test the efficacy of each policy combination.

[Code] [Paper] [Poster][Extended Abstract]

AI-Generated Art

The moral standing of robots and artificial intelligence (AI) systems has become a widely debated topic by normative research. This discussion, however, has primarily focused on those systems developed for social functions, e.g., social robots. Given the increasing interdependence of society with nonsocial machines, examining how existing normative claims could be extended to specific disrupted sectors, such as the art industry, has become imperative. 

Inspired by the proposals to ground machines’ moral status on social relations advanced by Gunkel and Coeckelbergh, this research presents online experiments (∑N = 448) that test whether and how interacting with AI-generated art affects the perceived moral standing of its creator, i.e., the AI-generative system. 

Our results indicate that assessing an AI system’s lack of mind could influence how people subsequently evaluate AI-generated art. We also find that the overvaluation of AI-generated images could negatively affect their creator’s perceived agency. Our experiments, however, did not suggest that interacting with AI-generated art has any significant effect on the perceived moral standing of the machine. These findings reveal that social-relational approaches to AI rights could be intertwined with property-based theses of moral standing. We shed light on how empirical studies can contribute to the AI and robot rights debate by revealing the public perception of this issue.

Lima, G., Zhunis, A., Manovich, L., & Cha, M. (2021). On the Social-Relational Moral Standing of AI: An Empirical Study Using AI-Generated Art. Frontiers in Robotics and AI, 8.

[Paper] [Data]


Political Personalization

Political scientists have long studied how the image of a nation is shaped in the media. Political personalization, the tendency for an individual leader to be the main focus of stories related to the nation, can impact its image, positively or negatively. Although not the first war in which social media have played a key role, the Russo-Ukrainian war is said to be the most viral. 

Using a Twitter dataset collected from the beginning of the conflict, we explore how political personalization is affected by the development of the war, with a focus on the role of sentiment. As expected, users' exposure to content, and the emotions expressed, play a role in shaping opinions of the war and its main actors. However, in the case of Putin, overexposure to negative content weakens this effect, which is likely due to desensitization. 

To our knowledge, this is the first large-scale study of political personalization, which models the dynamics between users' expressed emotions, and their exposure to consumed content during an ongoing war. 

This work is under review!🤞