🔸Human-AI Interaction: Artificial intelligence (AI) is rapidly being incorporated into our everyday lives. For many people, AI now functions as far more than an information search tool: students use it to study, researchers are increasingly incorporating it into their work (e.g., through content coding, stimuli generation, and AI-generated participants), and individuals facing mental and physical health challenges are increasingly seeking it out for advice. As the use of AI expands, so does the need to understand human-AI interactions. One central question in this space is how people perceive emotionally expressive AI. In human-human interactions, emotion expressions play an essential role in helping people infer the internal states and intentions of others. Imbuing emotions into AI may provide a similar benefit, making AI more relatable, predictable, and trustworthy. In turn, this may increase alignment between the human and AI which is important for the efficacy of future human-AI teams. The expression of emotion by AI, however, may not always have such a positive effect. Since AI cannot "feel" emotions in the same way that humans can, AI that expresses too much emotion or routinely mimics the emotions of humans may be perceived as less trustworthy or even manipulative. In emerging work, I plan to examine these questions to better understand the role of emotion in human-AI interactions.

🔸Mixed Emotions: Despite a wealth of evidence suggesting that people can feel positive and negative emotions (e.g., happiness and sadness) at the same time, little is known about these experiences. For one, outside of the laboratory, it is unclear what types of events occasion mixed emotions. In ongoing work, my colleagues and I are taking a two-stage, bottom-up approach to address this gap. In Stage 1, we used natural language processing (NLP) to identify the events that people associate with mixed emotions (e.g., major life transitions). In Stage 2, we are testing whether there is likely a causal link between these events and people's mixed emotional experiences. To do so, we are studying people's emotional experiences while they actually live through these events. For example, using a quasi-experimental design, we asked people to report their emotions as they experienced some of these events in real life (e.g., parents whose children are moving into college for the first time and college graduates on their graduation day). In addition, we have experimentally manipulated people's construals of these events (e.g., graduation) to better understand why these events lead people to feel mixed emotions. In addition to studying the events that occasion mixed emotions, this line of work also hopes to understand what mixed emotions mean for the individual. For example, are/when are mixed emotions beneficial/harmful for the individual's preparedness, functioning, and resilience?

🔸Disambiguating "Ambiguous" Facial Expressions: In a second line of work, my colleagues and I have been investigating two types of facial expressions that researchers have called "ambiguous", surprised and neutral faces. Prior theorizing suggests that surprise faces are ambiguous, because they can communicate that a pleasant (e.g., someone throws you a surprise birthday party) or an unpleasant event (e.g., a bear suddenly runs toward you) just occurred. Neutral faces, on the other hand, are thought to be ambiguous because they give little indication that anything has happened to the person at all. Despite these claims, perceptions of surprised and neutral faces share some similarities (e.g., people tend to perceive the people behind them both as feeling negative but also a little bit positive). In ongoing studies, we are investigating perceptions of surprised and neutral faces to better understand what makes them ambiguous, how people resolve their ambiguity, and the social consequences of ambiguity in emotional expression.

🔸Perceptions of Neutral Faces: One of our general findings from the line of inquiry above is that people tend to perceive the people behind neutral faces as feeling bad. Why might this be the case? In a related line of work, we are currently attempting to find out. By our analysis, there are at least two reasons for this neutral face bias. First, many people's neutral faces appear more morphologically similar to expressions of prototypical negative emotions (e.g., they often have downturned lips) than they do positive ones. Second, neutral faces may be perceived as negative because during social interactions, we often expect others to express some amount of positive affect or interest (e.g., there are display rules proscribing neutral faces during many social interactions). To investigate this possibility, we have been conducting a series of studies investigating (a) when display rules may prescribe and proscribe a neutral face and (b) how people perceive others with neutral faces as feeling in both types of situations. Generally, we have observed that people see neutral faces as negative.