As conversational AI systems become more integrated into everyday life, an important question remains: how do people actually experience these systems in real-world, culturally diverse contexts?
 

What if the biggest limitation in today’s AI systems isn’t the model, but the quality of the data we feed into it?