If it’s not Learning, it’s not Artificial Intelligence (AI) – it’s an Algorithm.
Before we dive in, I’d like to say upfront that this is by no means a definitive guide to Artificial Intelligence, Machine Learning, ELMs, Neural Nets et al. What this is intended to be is a simplification that serves to provide clarity on the key differences between Artificial Intelligence vs Algorithms – and help you identify which is which. Because right now, EVERYONE is claiming their app / website / analytics tool is AI. Some are. Most are not. Here’s how to tell the difference.
Let’s start with Artificial Intelligence
Definition of Artificial Intelligence
In computer science AI research is defined as the study of “intelligent agents“: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals. Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”.
source | Wikipedia
How Artificial Intelligence Works
Artificial Intelligence works like the human brain; through machine learning, a computational system runs millions of iterations of a sequence of events continuously until it is able to create a correct result with close to 100% accuracy and recurrence.
The intelligence comes from the ability to return results into the system, to the front of the routine, so that the next time it runs, it has new, previously non-existent data to add to the sequence. The AI system creates new data within itself – this is something algorithms cannot do.
Algorithms lack intelligence
Definition of Algorithm
In mathematics and computer science, an algorithm is an unambiguous specification of how to solve a class of problems. Algorithms can perform calculation, data processing and automated reasoning tasks.
source | Wikipedia
How Algorithms Work
An algorithm runs a sequence once, moving along a pre-programmed routine; If This, Then That – “If X = 100, then turn right”. Algorithms can feel like intelligence, but they are not. Algorithms lack the ability to “understand” whether an outcome is right or wrong (wanted or unwanted), they simply execute the sequence, over and over.
AI also executes sequences, BUT as those sequences get resolved into outcomes, the deep learning ability of an AI’s neural network reports back whether the outcome matches the desired result. For example, for the query, “is this an image of a cat?”, the answer is either Yes, or No – but the identification of cat vs non-cat criteria is complex.
An AI running millions of cat and non-cat images through its learning machine, is eventually able to identify the criteria for making a cat image. Just like a child learning maths, a correct answer is rewarded – and the machine looks to create more outcomes that match the successful criteria.
There is an ongoing process with AI that simply does not exist with algorithms. So please, stop saying your app is AI – when it is merely running algorithms. As smart as algorithms may be (and they can do a lot of cool stuff!), they are not Intelligent, so let’s stop calling them that.
Artificial Intelligence vs Algorithms; The Battle…
Algorithms generally follow a sequence of Yes/No options, until an outcome is achieved. Basically complex flow charts, they are linear and completed once the sequence finishes.
Algorithms can also match similarities in data sets to create a new data set of shared items (this is how recommendation engines work on stores like Amazon). Again though, this is not learning, its simply running a routine that looks for the same items on two lists, and creates a 3rd list that combines those shared items.
AIs on the other hand run many simultaneous sequences, mapping the outcomes against desired results and adjusting the sequence on-the-fly to create more outcomes that match the criteria of the desired result, i.e: they learn from trial and error what a desired result looks like as well as how to create as many of those outcomes as possible, as often as possible.
This is essentially where algos miss out, once their routines are completed they do not recycle that information back into the system to create a “better” result next time. they simply execute the same routine again.
Intelligent vs Smart
Hopefully this has helped identify the key differences between Artificial Intelligence vs Algorithms. Algorithms are exceptionally useful and can be very powerful – google is an algorithm – but they are more “smart” than “intelligent”. Most of the technologies that you see doing smart stuff, are running powerful algos (some with deep learning layers) that take bundles of data and sort them into useful results through a series of linear Yes/No type routines.
The Intelligence of AI is the ability to understand how a result is being generated and self-correct in order to achieve more desired results.
Artificial Intelligence vs Algorithms
So there you go, the next time someone tells you their app is “AI powered”, you’ll have the clout to question whether in fact it is actually running an Artificial Intelligence, or is simply running smart algorithm sequences.