Interjecting Into Conversations

One of the weird issues that’s going to come up over the next year or two is going to be how voice interactive devices are supposed to interrupt us. Today, these devices normally wait for us to speak a wake up word but it’s likely that they’re going to be pushing alerts or needing to interject.

There are two modes through which they could interrupt — blind or aware.

Blind Modes:

The Barge In. In this mode, the device just speaks the interruption. We did this with the Ubi and it can be very jarring.

The Chime. With this, the device makes a gentle sounds that then allows for the user to query for the message.

Aware Modes:

Presence Aware. The device knows that someone is there and listening, then either sends The Chime or The Barge In.

Interruption Aware. The device waits for a pause in the conversation and then alerts either with an announcement or an alert.

Time Aware. The devices knows that it’s too late or early to make a noise so it stays silent.

Context Aware. This is likely the most difficult. The device listens to the conversation, transcribes, and diarizes, then interjects at an important point. “Hmm hmm… actually, you’re both wrong…”