Picturing the Fictitious Person: User Perceptions of AI-generated Personas

Artificial intelligence has increasingly been leveraged to create personas in recent years, with tools like DALL-E and Midjourney generating detailed textual and visual representations. However, a recent study, Picturing the Fictitious Person: An Exploratory Study on the Effect of Images on User Perceptions of AI-generated Personas, has provided intriguing insights into the relative importance of textual content versus images in shaping user perceptions of these AI-generated personas.

The study’s primary objective was to determine whether the inclusion of images in AI-generated personas significantly influences user perceptions compared to personas without images. Surprisingly, the findings revealed that images, even when generated by advanced AI models, did not significantly impact user perceptions. Instead, textual content was identified as the primary driver of user evaluations. This suggests that the language, tone, and information embedded in the persona’s description are critical factors influencing how users perceive and interact with these digital entities.

The research highlighted several key aspects of persona-specific differences. Distinct personas, characterized by their unique content, elicited varying perceptions across multiple attributes. This finding underscores the importance of crafting persona content tailored to specific user needs and objectives. It emphasizes the necessity for persona creators to focus on the quality and accuracy of textual descriptions, ensuring they resonate well with the target audience.

Additionally, the study emphasized the impact of participants’ familiarity with personas on their perceptions. Those with prior knowledge or experience using personas exhibited more favorable perceptions, particularly in clarity, consistency, credibility, similarity, usability, and willingness to use (WTU). This highlights the importance of educating end users about personas and their applications, potentially through training sessions, workshops, or introductory briefings. Incorporating persona education strategies into design projects can enhance user engagement and satisfaction.

The research provides several practical implications for persona designers and highlights the importance of pilot testing personas before deployment. While images can enhance visual appeal, they should be viewed as complementary rather than pivotal elements. METRIC is one such tool that can be used for these pilot tests. By using eye tracking, persona designers can gain insights into which aspects of the AI design capture a user’s attention. An additional follow-up survey of the experience can grant them insights into exactly how the user felt about the designs (if the user found them disturbing, confusing, helpful, or which elements of the design triggered any of these emotions).

The study’s findings advocate for a strategic focus on textual content in AI-generated personas while recognizing the nuanced role of images and demographic factors. Persona designers can create more effective and user-centric digital representations by engaging end users in the evaluation process and incorporating educational strategies.

Try out METRIC now!

Read more research about the study below!

Salminen, J., Santos, J. M., Jung, S.G., and Jansen, B. J. (2024) Picturing the fictitious person: An exploratory study on the effect of images on user perceptions of AI-generated personas, Computers in Human Behavior: Artificial Humans, 100052. https://doi.org/10.1016/j.chbah.2024.100052

Leave a Comment