Google Assistant is… (or at least according to this article)
I’ve seen that chart reproduced in a few places, but the problem is that it appears that the original experiment ( not the publication of the results) was done before Amazon added contextual awareness to echo, which they did in December 2016. Prior to that time, you could set up a question list where echo would fail 90% of the time.
For example:
“Alexa, Who is the last President of the United States?”
“Alexa, Who is his wife?”
"Alexa, how tall is she? "
Will now correctly return Michelle Obama’s height. But it used to fail on the second question.
Always question statistical results that don’t give you the date of when the sampling was done. 
I REALLY wish the media understood the term AI…NONE of the systems in the consumer market are true AI…They are action/reaction…While the appearance is intelligence, every single response is pre-determined by some programmer somewhere. The voice recognition is SOMEWHAT AI if you give it feedback, and Google has the lead here…If you say a statement and it simply heard you wrong, clicking the accompanying app and telling it ‘this was wrong’ will help…but the failure here is that you don’t get an opportunity to tell it what is RIGHT. THAT is how a computer learns. Otherwise it is a very complex set of if/then statements.
I work in this field and I even cringe when someone internally (for our medical genome projects) starts talking about AI without having a full 360 degree feedback system where it can actually LEARN.
And don’t get me started about that movie AI 
Voice recognition by definition is artificial intelligence. That definition goes back at least to the 1980s, and probably before.
Learning is a different aspect of AI, but not the only one.
AI: the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
2 Likes
Yeah…my frustration is over the ‘generic’ AI used by people in the media that really think just a computer by itself is intelligent. The heuristic learning piece is really what I was referring to…
Ah…the 1980s…where we thought we knew everything a computer was capable of…I remember it well
1 Like
Asimov’s robots don’t define AI…A machine doesn’t have to be capable of “learning” on its own to be deemed intelligent. I work with IBM’s Watson and is nothing more than a complex set of if/then statements…
1 Like
Going back to my 1980’s reference, we know SO much more about what computers do that I think the debate here is whether we should actually DEFINE AI better…in 2017 terms.
Just think…in another 40 years we will be debating what the word ‘sentient’ '…and probably debating that WITH a robot or computer.
1 Like
In 40 years self driving cars will be called just cars and the cars that require a driver will be called antic cars. Our generation will still be arguing that they are not “cars” but self driving cars…
1 Like
You can freely create a Lambda function that allows Alexa to scrape the top response from a Google query. This is equivalent to self-publishing a SmartApp, but it gives you the ability to have Alexa ask Google for info. Best of both worlds. Profit!
2 Likes