Interactive Persona from Survey2Persona leveraging LLM for conversation
Here is how fast things are moving.
In April 2023, a crowd research company emailed out an April Fool’s Day message about using “simulated personalities” to replace humans in response-based research.
Our research team wrote a research article that this is not such a wild idea and that this approach would actually be a feasible methodology with good practical and research value. The article, Employing Large Language Models in Survey Research, was published in September 2023.
In November 2023, our research team’s survey system, Survey2Persona, introduced Interactive Personas, where a survey designer or decision-maker can converse directly with the personas to get in-depth and nuanced responses from their personas.
From analyzing the conversation, these Interactive Personas are making connections among the survey data that even a skilled survey designer might miss.
All this in about 7 months!
The next step is evolutionary.
The Interactive Personas can enhance the survey data with other data sources, such as social media data, other survey datasets, research articles, etc.
This is a method known as data triangulation.
If applied correctly, it can potentially greatly enhance the validity of individual surveys, resulting in better decisions and better results.
Read more!
Jansen, B. J., Jung, S.G., and Salminen, J. (2023) Employing Large Language Models in Survey Research. Natural Language Processing Journal. 4, 100020.
Turning Data-Driven Profiles Into Interactive Personas
https://s2p.qcri.org/blog/turning-data-driven-profiles-into-truly-interactive-personas/
