Are there actually different types of artificial intelligence available? That may on first glance sound a little strange; are we suggesting there might be a version that responds in rhyming couplets, or perhaps some platforms are simply not as clever as others. Incidentally, that is a tempting track to ride down and explore further at some point. But for now we’ll stay on this course and cover the two different approaches to AI solutions.
Essentially, AI is created and thus delivered using one of three approaches; Focused (or Narrow), General (or Strong)m and Super (or Incredible, Awesome or Scary, depending on your stance).
Let’s take a brief look at Focused Artificial Intelligence for this article.
Focused (Narrow) AI
For the vast majority of organisations, AI is looked at as a possible tool to use within help in explicit, defined areas. Let’s take a basic example of monitoring a kettle to give you information about its usage over time. To undertake this objective, the AI platform will need to handle a limited, set range of data. For a kettle this can include:
- The energy it requires to operate
- The amount of water it is holding
- The hours of the day it is used (and number of days per week)
- The ambient temperature of its surroundings
- The material the kettle is made from
- Even its height above sea level (if you want to be more accurate)
These data inputs are defined for the AI platform’s algorithms to undertake for you one or more set purpose. These could be the ideal time for the kettle to be first used in any given day, or when usage is potentially impacting its projected lifecycle. What you require from the AI platform is defined by the data it takes in and the set purpose of the algorithms in use. As you can see, we know what we want AI to be involved in and what it should help us with.
This is the reason this type of Artificial Intelligence is known as Narrow AI; it is applying Artificial Intelligence in very specific tasks It’s also known as Weak AI, which purely on a personal level we consider a little inaccurate as they potentially convey a false perception; the work AI can handle can be very detailed – the likes of autonomous vehicles or Google’s Alpha Go by definition use ‘weak’ AI as the data they utilise is defined in scope, yet you would hardly call such involved deployment of AI to be ‘weak’. So we tend to call this Focused AI, as it better describes the outward reason for using the technology, rather than the inward looking stance of ‘weak’.