Automation, Robotics, and Digital Twins

We began the Data Journey by stating that data is an integrated and inseparable part of our lives, society, business, industry, and the public sector. Data continually changes the rules of the game for how the world is connected. This is not just digitalisation, it is a digital transformation.

Even if you were to choose not to use any form of computer at all and only pay with cash, you would still depend on a staggering amount of data-driven processes in your daily life. Not directly, but as a secondary user.

Data is at the heart of it all, like oxygen for the algorithms that govern our digital everyday life.

We have talked less about this perspective on data—how it underlies digital technology, efficiency, and automation—in these last chapters. Therefore, let’s collect some threads here and be reminded of how data also creates value in other ways.

To “leave the job to the machines” is nothing new in itself.

Watermills and windmills are ancient technology. Since the industrial revolution towards the end of the 18th century, steam-driven engines and other industrial machines have performed tasks faster, better and on a larger scale than humans have been able to. Even programming with binary code goes back more than 200 years, as we remember with the use of punch cards in old industrial machines.

With the emergence of computers, the development of high-level programming languages, and technology such as artificial intelligence, we have been able to solve increasingly complex problems—with less and less human involvement.

The benefits are many: Automation can eliminate human errors and make processes more efficient, consistent, and reliable. It also facilitates a whole range of new products, services, and systems we use every single day.

Netflix recommends the films you watch, machine learning models recommend new employees, and robots replace manual jobs—both physically in factories and practically in customer service, for example. If you’ve seen a chatbot as the first line of defence on many company websites, it’s precisely because it’s so much more cost-effective to let a robot answer many of the most basic questions.

We also let Google and Apple fill out forms for us, our devices are automatically synchronised and we order items with one click. In short, we automate a lot of time-consuming, troublesome, and often downright boring tasks, so we no longer have to do ourselves.

But long before we automated digital processes, we automated physical processes. We also do this now in new and often much more effective ways, thanks to data—as the physical and digital merge in the Internet of things.

Much of the automation around us is driven by artificially intelligent systems. But far from everything.

Many chatbots, for instance, are not artificially intelligent at all, but rely on rule-based systems built on an “if A, then B” " instruction (similar to how decision trees work). They nonetheless make the banks, online stores and news media they “work” " for more efficient. (ChatGPT, admittedly, has raised the bar for what we expect from a chatbot now).

An example of this type of rule-based automation is what is called robotic process automation (RPA). You can read more about it in the accordion menu below:

A digital twin, as you may remember from the first chapters of the Data Journey, is a digital representation of a specific system or object.

This type of digital twin is not something most people work with directly, but here too we are in many ways “secondary users”, as this is used for example to ensure effective power supply or build complex motorway bridges.

Say, for example, we want to explore a project for offshore wind. Based on data about the wind turbines themselves, and the environment they are part of—the sea and the coast—we can use machine learning models here to simulate and predict how the turbines and the surroundings will affect each other, how much power will be generated, and so forth. This is a form of digital twin.

We can also construct the entire offshore wind park in a game engine—like a video game—which future employees can explore in virtual reality (VR). With VR glasses on, they can get the experience of being out there before the turbines are even constructed, and gain training and experience with what it will actually be like to work there. This too is a form of digital twin.

This can also be connected to real data about, for example, weather and wind—and real-time data from the turbines themselves after they are constructed.

Robotics is the science behind building robots. With robotics, we usually think of physical robots, as opposed to digital software robots as in the case with RPA.

However, these don’t necessarily have to be humanoid creations. Robotics, more generally, is about constructions/physical systems that support humans and help us solve problems.

A robot can, for instance, be used to sort goods in a warehouse, using actuators (“arms”) and sensors (“eyes and ears”) that allow the robot to register and interact with its physical surroundings.

This is also very much about the use of data. For the robot to be able to do anything, it must register data about its surroundings, which is digitised (turned into 0 and 1—bits and bytes) and processed, to then trigger some form of action or response.

Some robots are “dumb”, i.e., programmed to perform simple, repetitive tasks with no form of AI. Others can register and react to their surroundings in ways we perceive as more or less intelligent.

This applies, for example, to self-driving cars and robotic vacuum cleaners, which rely on machine learning, and which can learn, adapt, and improve over time through software updates and machine learning models, without the physical hardware needing to change.

This is of course closely tied to automation—which is becoming increasingly widespread in both the physical and digital world. This will inevitably have consequences for how we live and work, as more and more tasks can be taken over by machines.

However, such systems cannot in any way feel, think or reason. Precisely for this reason, they cannot currently replace tasks that require creativity, collaboration, intuition or empathy. They are also bad at handling exceptions that in some way deviate from what they are programmed to do, and this is where we humans still have an ace up our sleeve.

Because no one beats us at flexibility and adaptability. At least not yet.