The Jibo bot is at CES 2017, but it is strictly in demo mode.

I tried speaking to one, waving my hands before someone at the booth explained that it wasn’t working yet.

As you may know, this service bot is intended for families and even at the office. It can, according to the site for the bot, MC a game with kids, detect who is talking and face that direction, and react with emotion. The bot at CES didn’t do any of those things. It was in a dance routine, happily ignoring everyone around it.

According to many reports from this fall, Jibo suffers from a latency problem, which is a broad way of saying that the bot doesn’t always work on a home router over Wi-Fi. More importantly, the latency is a more general term that means the digital connections, sensors, and microphones get easily confused. Another bot called Wavebot was also in patrol mode, gliding from one pre-set place to another. A rep in the booth explained that there are too many signals, too many voices, too much interference. Wavebot and Jibo would have been easily confused and intercept too many signals.

AI Weekly

The must-read newsletter for AI and Big Data industry written by Khari Johnson, Kyle Wiggers, and Seth Colaner.

Included with VentureBeat Insider and VentureBeat VIP memberships.

This is a bigger problem than any of us might realize.

Jibo is just one example, but digital connections are everywhere. In my home, which is a bit like being in a test lab, there are Bluetooth keyboards everywhere, a security system that can open and shut my garage doors, plants that use Wi-Fi to tell me when they need water, and a simple bot called the Cozmo that can play games with my kids.

Add a semi-sentient being into the mix? You’re asking for trouble. Add a chatbot or voice assistant that tries to listen to the room? In my home, that doesn’t work either. After moving the Google Home speaker into my living room last week, I realized the Google Assistant would not work there. For one, it hears OK Google during commercials constantly. More importantly, it can’t hear me when anyone else is around. Amazon Alexa has a good microphone, but it also gets confused by ambient noise. For anyone in robotics, making a chatbot, or creating an AI — it’s time to think about this topic.

Jibo is just one example, but it is a good one. We want social bots who can react, ask questions — think differently. The promise with Jibo is that it will be like another family member, but this particular digital pet might not know the difference between a commercial and your own voice, it seems. It might work if you sit up right in front of it and talk slowly, but bots need to be in the background, listening to us and interjecting ideas once in awhile.

Plus, the interference is getting worse. I have doorbells that connect, my car can use the Wi-Fi in my house, the Apple AirPods I used all day emit a signal. Bots will need a way to communicate reliably.

That could mean creating a new standard — one that is only for bots. It could mean using complex AI that knows when I’m talking to the bot and not talking about the bot. It might mean limiting some features to make sure the bot always works.

I’m curious to hear what you think — is latency really a big problem? How can bots adjust?

At least for now, the Jibo bot I tried to test wasn’t ready for prime time.

CES2017