Study: AI-Enhanced Smiles in Video Chats Affect Dating Interest
A recent study from the University of Glasgow has demonstrated that AI-driven smile-enhancement filters can subtly alter the romantic attraction felt between individuals in real-time video interactions. The study, published in Proceedings of the National Academy of Sciences, used AI to manipulate participants’ smiles during video speed-dating sessions, either by increasing or decreasing the size of their smiles without their knowledge.
During these sessions, researchers adjusted smiles to either align or misalign between partners. Findings revealed that aligned smiles led to higher attraction ratings between participants than those without synchronized smiles. Notably, this increase in attraction occurred even though participants did not realize their smiles were being altered, showing how nonverbal cues, such as facial expressions, influence interpersonal attraction.
In addition to smile alignment’s impact on attraction, the manipulated smiles also influenced vocal patterns and participants’ propensity to mirror each other’s expressions. Interestingly, participants believed their date was more attracted to them when AI-enhanced smiles were used – if both were using the filter, then they would presumably create a feedback loop of attraction.
With filters already being a common thing on dating sites as a whole, and the prevalence of these real-time filters on video platforms, there is a very high chance that more dating sites begin to adopt features like this in the long term. Video chats provide greater authenticity in the same way that voice notes do, and could be one of the solutions to the current demand for authentic and deeply personal dating experiences – with these filters potentially being a factor as well.
Lead researcher Dr. Pablo Arias-Sarah emphasized that this study raises ethical questions about the future of AI-driven filters, especially as they become more realistic. Ethical guidelines may be needed to regulate their influence on social behaviors, similar to the recent discussions around deepfake technology.
