Persuasiveness of Conversational Agents for Targeted Advertising: Autism and Gen-AI Chatbots

Kirsten Chapman, Kaitlyn Klabacka, Garrett Smith, and Xinru Page

Marketing firms are starting to leverage Generative-AI-based chatbots as a more persuasive form of advertising. This is potentially more harmful for autistic young adults with substantial support needs who tend to have lower financial literacy, be more trusting of bad actors when communicating through social technologies, and more likely to anthropomorphize objects. Thus, we investigated whether a chatbot assuming the likeness of a favorite celebrity would more greatly influence consumer preferences of this population when compared against either autistic individuals without substantial support needs or the general population. We conducted an experimental survey where participants 1) ranked their preferences of various consumer items, 2) performed intervening tasks answering several questions, 3) viewed a chatbot interaction where the chatbot exhibited opposite consumer preferences, and then 4) re-ranked their preferences for those items. We found that autistic young adults with substantial support needs were more likely to change their preferences than the sample from the general population. Moreover, they were also more likely to change their preferences than autistic young adults without substantial support needs. These findings suggest that autistic individuals with substantial support needs are more susceptible to celebrity chatbot persuasion. We discuss the risks and guardrails that need to be associated with deploying generative-AI-based chatbots for targeted advertising.

A House Divided: How U.S. Politics Could Shape Contact-Tracing Adoption in Future Pandemics

Garrett Smith, Kirsten Chapman, Tzu-Yu Weng, Haijing Hao, Mainack Mondal, Staci Smith, Yunan Chen, and Xinru Page

Contact tracing has shown to be an effective tool in limiting the spread of transmittable diseases in countries where it is widely adopted. During the COVID-19 pandemic, contact tracing app adoption in the United States was low despite having the highest number of recorded cases worldwide. To better understand why, we conducted a survey (N=302, matched to U.S. census demographics) and found that political orientation overwhelmingly predicted attitudes towards COVID-19 and the adoption of contact tracing apps. These attitudes also overwhelmingly shaped people’s willingness to participate in contact tracing for diseases in future pandemics. Our findings suggest that the politically charged environment surrounding COVID-19 in the U.S. may have a long-term impact on American’s willingness to utilize contact tracing for diseases in future pandemics. We conclude with recommendations for technology designers and policymakers on how to overcome the sharp divide that has been driven by the political discourse in the U.S.

A Privacy Paradox? Impact of Privacy Concerns on Willingness to Disclose COVID-19 Health Status in the United States

Kirsten Chapman, Melanie Klimes, Braden Wellman, Garrett Smith, Mainack Mondal, Staci Smith, Yunan Chen, Haijing Hao, and Xinru Page

Privacy concerns around sharing personal health information are frequently cited as hindering COVID-19 contact tracing app adoption. We conducted a nationally representative survey of 304 adults in the United States to investigate their attitudes towards sharing two types of COVID-19 health status (COVID-19 Diagnosis, Exposure to COVID-19) with three different audiences (Anyone, Frequent Contacts, Occasional Contacts). Using the Internet User’s Information Privacy Concern (IUIPC) scale, we were able to identify the effect of different types of privacy concerns on sharing this information with various audiences. We found that privacy concerns around data Collection predicted lower willingness to share either type of health status to all of these audiences. However, desire for Control and for Awareness of data practices increased willingness to share health information with certain audiences. We discuss the implications of our findings.

"Others Have the Right to Know": Determinants of Willingness to Share COVID-19-Related Health Symptoms

Kirsten Chapman, Melanie Klimes, Braden Wellman, Garrett Smith, Madeline Bonham, Yunan Chen, Staci Smith, Mainack Mondal, Xinru Page

Consideration for data privacy is a potentially significant factor behind user-apprehension regarding sharing personal medical information (e.g., disease symptoms), even during the COVID-19 pandemic. In this study, we set out to unpack the extent to which privacy (and related) factors influence people’s perceptions of data sharing. Specifically, we designed and deployed a survey (N=304) with both qualitative and quantitative questions concerning willingness to share COVID-19 related medical information with others. Our findings indicate that although an individual might feel strongly about maintaining their privacy in general, in context of a global pandemic, altruism may outweigh these concerns. This may be especially true for others who are frequent contacts. We discuss the implications of societal duty having a stronger effect on willingness to disclose medical information than privacy amidst a pandemic.

Privacy Interventions and Education (PIE): Encouraging Privacy Protective Behavioral Change Online

Garrett Smith, Kirsten Chapman, Zainab Agha, Janet Ruppert, Spring Cullen, Sushmita Khan, Bart Knijnenburg, Jessica Vitak, Priya C. Kumar, Pamela J. Wisniewski, and Xinru Page