Jibo Emergency Response Demo


I created an app showing how Jibo might respond to verbal indications of a potential emergency situation such as fire or smoke or a request for help. If someone tells Jibo “I smell smoke”, “fire”, “I need help”, “I’m hurt”, etc., Jibo will inquire if help is needed and notify the appropriate authority if the response is yes. Additionally, in situations that could potentially involve a bad actor, Jibo videotapes the scene and asks for a password before returning to a non-emergency state. In all cases, Jibo also emits an alarm.

You can view the demo here:

Jibo Emergency Response System

You’ve dmonstrated a very important use case here which I think would be critical for Jibo to have if he is in any caretaker-type of capacity.

Per your demo, I don’t necessarily agree with getting forced into contacting the authorities if you don’t have the correct password. For instance, I can imagine a child asking Jibo, “Hey Jibo, I lost my teddy bear. Help me please” which would trigger your emergency response. If the child doesn’t have the password, and they probably shouldn’t, then the police end up getting contacted for something trivial. Have you considered going the other direction with it and requiring a password to actually contact the police/fire dept.?


First of all, GREAT WORK! This is a very important topic. There are similar discussions going on right now for Echo, Google Voice and Siri. I encourage you to search and read up on how these assistance are addressing the issue and then see if you can innovate from there. This will give you the benefit of their studies and conclusions.

I do understand the password bit. I think that I personally would use that method but not with simple help requests. Rather I would have a duress phrase. And I would absolutely have that be password protected.


I know Jibo is meant to recognize faces, but does it recognize voices alone? This might help with some of the password issues, although in an emergency I probably squeal like a 4 year old girl.


Yes, the ASR (Audio Speech Recognition) module returns a speakerName that it determines from the voice, so no face needed. Obviously, though, if you’re squealing like a 4-year old girl, Jibo might have a hard time figuring out who you are. :slight_smile:

Record your skills on a live Jibo!