Right now Amazon Echo Alexa is the question and answer system. You ask a question “Who is George Washington?” and "she" responds, usually by scraping Wikipedia for the answer, and nicely answer your question. Nothing is wrong with scraping Wikipedia, in fact, it's awesome!! You don't have to go to a computer and navigate through the user interface to get the answer.
What I want in these and products like it (including ours) is to have the device ask us questions when appropriate. Humans created Wikipedia, it’s maintained by humans, and edited by humans. That concept can be extended with our product and Alexa by allowing facts to be created and edited on the device. People, not everyone, would be more than willing and happy help educate Alexa/Our product If they knew answers to questions.
As an example, if I said a fact “ a cat is a small mammal etc..". In an empty knowledge system where ‘cat’ has not been defined the statement would then define ‘cat’. This is awesome and key to what we are doing. But, “mammal” would potentially not be defined at the time. It would make sense to ask the human “I don’t know mammal, can you define mammal please?”. The user could, of course, say “no” or “don’t ever ask me to define anything” if they wanted or they could define it. Obviously, when and how often the questions are asked depends on perhaps the type of customer but is certainly valid. With potentially millions and millions of people using these products the amount of information that could be added is enormous. And, as it is added with only voice, ANYONE can add it, ANYONE!
Having voice-only interaction limits how long timewise a question or answer can be. This means that we can have small definable pieces that people can readily answer/ask. We're not expecting people to write or say Wikipedia articles, reformat them appropriately, etc... We’re saying give us a quick definition. Many times when going to Wikipedia people are not interested in the whole article on ‘cat’ anyway, just a quick portion of it.
The interaction here is two-way. Computer-Human with voice. There is no complex computer mechanism ‘wikipedia’ to make us humans suffer to add information. Like Wikipedia however, there should be validation methods to guarantee people are not crazy or inaccurate with definitions. I’m confident there will be ways to do that and considering people's voices are recorded and potentially played back they will be more honest.
I look forward to our product and others asking people questions about the world.