Apple already has Siri as an AI. Now, it will also have Google Assistant as an AI. Or at least, the user would have the option to install and run Google Assistant. This is the massive undertaking with the user being the default winner.
Data and More Data
As an AI, Siri is used available user data and internet access to come up with correct and proper answers asked by users. This is the reason why it may run into wrong answers sometimes. If it latches on to a site with the wrong information, that information would be passed to the user regardless. Suffice to say; it does not have filters to differentiate what is correct and what is not. The data from Siri all comes from the iPhone, or whatever Apple iOS device it runs on.
In the same token, if you ask Siri a question, it is as if you asked a search engine. However, instead of coming up with a list of possible sites to check for answers, Siri provides the most probable answer. This sometimes results in silly or outright funny answers. However, it also results in the totally unacceptable result. No matter what Siri’s answer is, it is still up to the user to evaluate the answer and not take it at its face value.
Not Always Correct
If you fall in love with Siri’s answers which are really like magic, you should be afraid of the next thing that can happen to you. You would take any Siri answer at face value, even if it is wrong. This is similar to a taxi driver who uses a GPS system to navigate from Point A to Point B, a hotel for instance. There would be days when the GPS would say that you have arrived at your destination, which is the hotel, even if you are short ten meters, and not even on the driveway. The taxi driver can say, that you have arrived at your destination because it says so on the GPS. Besides, you can see the hotel entrance ten meters away up the driveway.
AI and GPS programs are not always correct. It all depends on what data it uses. The wrong data will always lead to embarrassing results. This is true not only of Siri, but also of any other AI, including Microsoft’s Cortana, Amazon’s Alexa, and Google Assistant.
Google Assistant on the Apple iPhone
Google is now planning a release of its AI, Assistant, on the Apple iPhone. This is great news for all concerned. But maybe not for Apple or for Siri. Although admittedly, Siri needs all the competition it can get so that it would not feel isolated in its own backyard.
With the choice of AI, users can now test and evaluate for themselves whether they would want to use Google Assistant or Siri. This has not happened before. However, this can also spur Google to make Assistant even better. The main advantage of any AI is the mountain of data it brings to the competition. In the case of Google that is a lot. They have more data about a person than any other company. Although it is possible that Facebook would have data on more people than what Google has. It should be noted that Facebook has its own machine learning software, but it has not released any AI. The machine learning software is responsible for Facebook Graph, which has not been widely released as yet.
Google Assistant on the iPhone does not sound right. However, it is a reality which should have been obvious a long time ago. It just so happens that Google was not interested in it until lately after they have introduced Google Assistant to the market.
On the iPhone, Assistant can access data that is on the iPhone. It can also pull data from Google Searches, as well as from Google apps like Photos, Plus, Drive, search, YouTube, and others. There is a mountain of data which the AI has to dive into to understand a single individual.
On top of that, Google Assistant can also access whatever data is on the iPhone. This is not just limited to the contacts and address book, but also other data that is already on the iPhone. Like Siri, Assistant thrives on data. With more data, it can churn out better information.
Filtering Wrong Information
One problem developers have encountered in recent years, is the proliferation of fake news. This is a real threat, and websites like Google and Facebook are trying to actively find a solution against fake news. Fake news is not malicious per se. However, it has been used instead, and then the references to itself are used to make people believe that it is true. The closest understandable explanation is that of a guy who spreads rumors and false information. After he releases the information, he forgets about it, and then when he hears about it from two independent sources, he praises these news sources for giving true data
This is the kind of information which current AI is afraid of. It is also afraid of somehow making it worse. The reason for this is that AI, as of the present, does not annotate news articles. Data has to be verified. It is no longer needed to have two independent sources to prove the truth. It is also necessary to put things into context when reporting about it.
