Written by Jareef Martuza, Ph.D. Candidate at NHH
Credit: Image created by Deniz Lefkeli using AI with https://pixlr.com/tr/.
We know our attitudes shape behavior in important ways. But sometimes, the behavior itself can shape attitudes. In their latest research “Sharing information with AI (versus a human) impairs brand trust: The role of audience size inferences and sense of exploitation” forthcoming in the IJRM, Deniz Lefkeli, Mustafa Karataş, and Zeynep Gürhan-Canli conducted several experiments to find that consumers trust brands less when they disclose information to AI as opposed to humans.
Why? What does this imply? I had the opportunity to discuss these and more with the lead author Deniz Lefkeli.
Research in the time of COVID-19
Most of us do not look back so fondly to when COVID-19 was still prominent, but the motivation for new research can come from the strangest of places.
“Face-to-face interactions with service employees were limited, so firms relied more on Artificial Intelligence (AI) in collecting customer information. This made us wonder how sharing information with AI might influence consumer trust in the brand.”
-Deniz Lefkeli
This question is important because consumers may often need to share information with firms before any sort of relationship is formed. In such cases, each touch point with the firm becomes a building block of trust, or conversely, erodes trust. With this in mind, Deniz and her then-dissertation advisor Zeynep Gürhan-Canli started this project where they studied the effect of mere information sharing on consumer trust. Subsequently, Mustafa Karataş, who shared the same advisor as Deniz and so an “academic brother”, joined this pursuit.
Although now spread across Rome, Astana, and Istanbul, all three authors were based at Koç University when conducting their research.
What, why, and when
Across nine studies, including A/B tests on Google Ads and online experiments, the team found that disclosing personal information to AI, compared to a human agent, leads to lower levels of brand trust. Specifically, people were less likely to click on ads that prompted sharing information with AI (vs. person), and they reported lower brand trust across different scenarios.
It is the “why” part that surprised the authors. “We were intrigued by how information disclosure led consumers to consider audience size and potential uses of their data. People will infer what will happen beyond the medium they share information with. They have a lay belief a big audience may be privy to what they share with an AI.”
Indeed, the authors found robust evidence that inferences of bigger audience size heightened consumers' sense of being exploited, which in turn lowered their trust in the brand. Further, the link between the sense of exploitation and brand trust was especially stronger among consumers with higher privacy concerns.
So what?
Deniz and her colleagues went beyond solving the theoretical puzzle. In subsequent studies, the authors found that explicitly informing consumers that their shared information is confidential and giving AI human traits can reduce the negative impact on brand trust.
“It is crucial to communicate that whatever information consumer shares, it is confidential, and its usage is restricted to a limited audience. This can enhance the desirability of sharing data, which can be a win-win for brand outcomes and consumer welfare.”
-Deniz Lefkeli
Read the paper
Interested in delving deeper into information sharing and trust? Click here for the full paper!
Reference
Lefkeli, D., Karataş, M., & Gürhan-Canli, Z. (2023). Sharing information with AI (versus a human) impairs brand trust: The role of audience size inferences and sense of exploitation. International Journal of Research in Marketing, S0167811623000654. https://doi.org/10.1016/j.ijresmar.2023.08.011
Meet Deniz Lefkeli
Assistant Professor of Marketing at LUISS Guido Carli University
How has your research influenced your opinion?
My research on AI and privacy and thereabouts influenced my views on how adaptable consumers are in adopting technology. It highlighted the need for empathy towards consumers who are less familiar with technology and the importance of informing them about data practices.
What are your current and near-future research goals?
My primary goal for now is to enhance consumer well-being in their interactions with technology, with a particular focus on privacy concerns and the varying roles AI can play.
Had you been in research and academia, what would you be?
I once had dreams of becoming a pilot. I would enjoy being in the sky. I think I am also drawn by a sense of freedom, responsibility, and control, and family admiration of pilots as well.
What’s your advice for young researchers?
Pursue research topics that you have a genuine interest in. For me, this led to exploring novel areas like AI, privacy, and information disclosure, which are some of the puzzles that drive me to do the work I do.
This article was written by
Jareef Bin Martuza
Ph.D. Candidate at the Norwegian School of Economics
