When Is the Best Time for an AI Speaker to Talk?

Data Analysis Results for Developing Optimal Utterance Timing Inference Technology
Returning Home, Smartphone Use, and Housework Are Good Times for Conversation

[Asia Economy Reporter Junho Hwang] When is the best time to talk to an artificial intelligence (AI) speaker?

A research team led by Professor Lee Euijin of the Department of Computer Science at the Korea Advanced Institute of Science and Technology (KAIST) installed AI speakers in the rooms of 40 students (2 per room) and collected a total of 3,500 user response data points over one week to find the answer to this question, the team announced on the 28th.

The research team identified major situational context factors for determining the appropriate timing as personal factors, movement factors, and social factors.

Personal factors include four main aspects: activity concentration, urgency and busyness level, mental and physical condition, and the possibility of listening or speaking for multitasking. For example, it was difficult to have a conversation with the speaker when the user was concentrating on studying or drying their hair with a hairdryer. Movement factors include three types: going out, returning home, and activity transitions.

In particular, the distance at which the user can converse with the AI speaker during movement greatly influenced the judgment of the optimal timing. Going out refers to moving outside the conversational range of the AI speaker, while returning home refers to moving back within the range. The returning home situation, where the user moves into the range, was mostly classified as a good time to talk.

The research team also discovered that not only phone conversations but also the presence of someone else affects the good timing for talking with the AI speaker. This is because users want to minimize conflicts caused by conversations with the AI speaker when their roommates are sleeping or concentrating on some activity.

Chana Rae, a student who participated as the first author in this study, said, "This research will be an important foundation for the development of future smart speakers," adding, "In the future, by utilizing situational context information detected through sensor data, AI speakers will be able to proactively detect the best timing to start, stop, or resume conversations and provide intelligent voice services." The results of this study were recently published in the international journal in the field of ubiquitous computing, Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies.

This content was produced with the assistance of AI translation services.