An A-Z of AI and automation



AI and automation are related technologies. But they’re not the same thing. Short for artificial intelligence, AI covers tech tools that mimic human intelligence and ability. Automation, meanwhile, follows human-made rules to make tasks, processes and workflows happen automatically.

These technologies also embody a field rampant with terminology.

With that in mind, here’s an A-Z of some of the biggest terms and fields surrounding AI, automation, or both.


Algorithm

An algorithm is a step-by-step guide that tells a computer how to do something. They’re a core component of computing — including AI and automation.


Big data

Big data is a buzzword that denotes extremely large data sets.

Automation is a core tool that helps companies collect, organise, transform and store big data. Data also plays a huge role in the training of different AI tools. In turn, these AI tools may then analyse and draw conclusions from big data, unlocking the value of raw data.


Chatbots

Chatbots are computer programs that chat with users. They’re closely related to AI and automation. This is because chatbots offer a form of automated conversation.

Many chatbots work by following pre-defined rules (like automation). As they’ve evolved, they’re starting to adopt AI capabilities and better emulate human conversation.


Disruption

Disruption refers to the large-scale upheaval and change that AI and automation bring. As these technologies settle into more tasks and roles, it’s on us to embrace the disruption as a transition phase on the way to the future of technology.


Ethics

With the change and disruption that they bring, AI and automation inspire ethical questions and concerns. For automation, most ethical discussions surround the impact of automation-fuelled job loss.

For AI, ethical discussions permeate every area of the technology. There are questions of accountability, concerns of algorithmic bias, and challenges with transparency. All these serve to inform wider AI acceptance.  


Future

AI and automation undoubtedly make up a core thread in the fabric of our future. It’s not a question of if these tools will impact the future, but how.


GTD

GTD is short for ‘getting things done’. It’s a method that helps you tackle and achieve your goals. It’s also a great way to approach your business process automation deployment.  


Hidden layer

Hidden layers are part of an artificial neural network — a computer system that simulates a biological brain. They’re a core part of deep learning (a subsection of AI). Hidden layers are layers of processing that we can’t see. So, we can’t know how the computer is analysing the data. We only see the input and output.


IF statements

IF statements are conditional statements that tell a computer what to do with certain information. They outline both what to look for, and what the system should do if a condition is (or isn’t) met. With automation software, you create IF rules that tell the software what to automate, and how.


Jobs

When discussing AI and automation for long enough, you’ll invariably encounter the assertion that ‘robots are taking our jobs’. While AI and automation stand to absorb and change some jobs, the technology will also create new jobs and new responsibilities.


Knowledge-based system

A knowledge-based system is one that uses an inference engine and a knowledge base to solve problems and draw conclusions. This means that the system applies logical rules to provided facts about the world to deduce new information.


Labelled data

When training AI, you need a lot of data. In some cases, this training data is labelled. This means that the data has meaningful tags that tell the machine the properties and answers to look for when viewing similar data. In other words, it’s data that comes with the answers that you want the machine to learn.


Machine learning

Machine learning (ML) is a subset of AI that focuses on enabling machines to learn from data and experience. There are many different types of ML. For example, deep learning, transfer learning and reinforcement learning.


NLP

NLP is shorthand for ‘natural language processing’. It’s another area of artificial intelligence. Its focus is to allow machines to understand and use human language as it’s spoken naturally.


Over-reliance

Both AI and automation use share a common pitfall: over-reliance. In either case, relying too heavily on either technology can cause untold issues. Both technologies are not infallible — they can run into unforeseen errors.


Productivity

A productivity boost is a benefit that both AI and automation offer.

Automation software, for instance, stands to boost productivity by removing the hurdles of shallow work. (Think admin, data entry, etc.) Meanwhile, AI increases productivity by supporting the decision-making process.


Qualification problem

The qualification problem is the impossibility of identifying and outlining all possible preconditions (and exceptions) that should lead to a given outcome.

In simple terms, a programmer cannot provide every possible message a free user might send when looking for a given answer. This is where AI, machine learning and NLP might be the answer to fill the gaps.


Rule-based system

A rule-based system is where a computer applies human-made rules. These rules allow it to manipulate, organise and store data input. Automation is an example of a rule-based system. 


Superintelligence

There are different levels of AI ability. Currently, we have narrow AI — good at one specific thing or area. We’re striving to reach general AI, which would understand, think and act at the same level as a human being across all subjects.

Then, there’s superintelligence, which is AI that has surpassed human understanding and ability. For now, superintelligence only exists in the realms of science fiction.


Turing test

The Turing test is a staple part of AI history. Devised by Alan Turing, the test examines a computer’s ability to show intelligence that’s indistinguishable from a human’s.


Unlabelled data

Unlabelled data is used for unsupervised machine learning. It’s data that doesn’t have any meaningful tags or answers attached to help an AI draw the right conclusions. Instead, the machine must notice patterns on its own to generate output.


Vision

While vision is something that comes naturally to humans, it’s much harder to teach a computer to ‘see’. Computer vision is another subset of artificial intelligence. It’s concerned with teaching computers to understand the visual world.

It allows machines to identify the subject of pictures and forms a large part of the technology behind facial recognition.


Winter

Even though the forecast for AI and automation predicts sunshine and clear skies now, this wasn’t always the case. AI has suffered at least one winter period. This is where faith in the technology plummeted — meaning less funding and less advancement.


(E)xpectations

For automation, over-hyped expectations lead to the ‘golden hammer’ approach. That is, the belief that the tool can fix everything and replace your human team. This, predictably, leads to disappointment.

For AI, over-hyped expectations can result in disillusionment and a loss of confidence in the tech. Over time, this could result in another AI winter.


Your processes

Automation software isn’t an out-of-the-box solution. Instead, it’s a platform, meaning that you can choose what to automate and how. It’s your processes and your programs for your business.


Zeroes and ones

What better way to end this A-Z of AI and automation than with the core of it all?

As smart as AI gets, as useful as automation proves, the technology — much like any computing ability — all boils down to ones and zeroes. Without the binary code that lets computers do what they do, life today would be very different.


An A-Z

This A to Z doesn’t encompass all the terms and discussions that revolve around AI and automation technology. But hopefully, it serves as a handy glimpse into the world of these technologies.

Are there other AI and automation terms that interest you? Tweet us and let us know!


Useful links

Are AI ethics impossible?

Robots taking our jobs: the four ‘D’s to factor in

Deep work and automation