There’s a lot of talk about how AI will support work and change the job market. I’ve already dedicated some space to this issue. There’s a certain trend with the potential to change our work and personal lives. Today, you will read more about the evolution of AI towards our personal assistant, and eventually, perhaps even a digital twin.
First… I’ll admit to a theft!
A digital twin, a popular buzzword, refers to a virtual copy of a real-world entity or a process or system. In fact, it could be a virtual copy of anything, widely used to perform simulations, tests, or other operations even before the real object, which the digital twin is meant to reflect, comes into existence.
I must admit, I like the term “digital twin”, but not so much its definition. So, lacking a better solution, I’ll borrow this phrase and define it as follows:
A digital twin is an AI-managed solution aimed at relieving us (humans :)) in performing certain tasks or making decisions based on commands, algorithms, and/or data. It can perform operations under our supervision or autonomously.
Now that the confessions are out of the way, let’s get to the point!
AI as our companion will increasingly relieve us of tasks where we lack knowledge and skills, while also handling those that are not particularly challenging. This should limit apathy and boredom at work. Meanwhile, we will gain more time for tasks that put us in a state of flow, where we have great competence and which are an engaging challenge. The state of flow, in turn, brings satisfaction from work and a sense of fulfilment.
Sounds beautiful, doesn’t it? After all, AI (especially Gen AI, generative artificial intelligence) is a technology of unlimited patience, using a powerful knowledge base for its operation. To some extent, I believe this vision is realistic, though it requires further development of technology and our competencies related to exploiting this potential.
This mutual development will form (and in part already forms) the basis for treating Gen AI as our virtual assistant, especially if we add various work automation tools. However, I think AI will become something more – it will provide increasing support in daily life and in building our wealth. (At the end of the text, I’ll encourage you to read about my 3W Navigation concept.)
Here we come to the concept of a digital twin.
Our Digital Ego
How, in my understanding, does a virtual assistant differ from a digital twin? A virtual assistant executes our commands, while a digital twin acts more autonomously, knowing us and the world around us. (The digital twin should not be confused with an avatar in the metaverse, though potentially a digital twin could take such a form.)
This is not as revolutionary as it might seem at first glance. We already do a lot under the dictates of algorithms that encourage us to move, remind us about water, and set aside the change for some foundation. But this is just a modest beginning.
One could risk saying that the concept of a digital twin operating thanks to AI fits into speculations about the upcoming era of hyper-personalization (i.e., highly advanced personalization of content and experiences). Hyper-personalization is possible, among other things, thanks to AI, which can deliver well-profiled materials for everyone. But! In my opinion, hyper-personalization will also be related to what our digital twins will be based on and react to. This could mean an interesting spiral of messages (conversation between AIs).
Let’s look at this issue from a marketing perspective. Broadly speaking, marketing today is a complex field that, using various channels and based on advanced algorithms (including AI), tries to convince us to take some action. Supposedly, in the future, a digital twin will relieve us in certain actions and decisions, and perhaps this will open a new field of marketing: marketing targeted to AI.
Soon, I would like to share with you, I hope, an interesting concept that, in a way, is a consequence of what I wrote above, so I invite you to subscribe to the newsletter – the subscription form is located under the article.
A digital twin can influence what we do and how we think to a greater extent than any other technology and concept implemented so far. At the same time, it can be our source of knowledge and wisdom (some futurologists say that just as we stream video, we will stream intelligence).
This creates fantastic opportunities and has the potential to make the world more equal than ever before. But it could also do the opposite: create a new technological barrier (though some costs must be incurred, including the need for infrastructure), and lead to changes in our behaviour, including the unlearning of how to learn and make decisions with an awareness of their consequences. Thus, the digital twin as a companion in making good choices and better understanding the world will be an opportunity for those who will want and be able to take advantage of it.
Digital Twin in Practice
Let’s consider how the coexistence of us and our digital twin (or maybe there will be many of them?) might look in practice. I don’t think a great revolution is needed, as the path has already been paved; it’s just a matter of properly combining elements.
In my opinion, functionalities related to AI as our personal assistant will be combined with a profile, within which information about us is collected from various devices. This could be, for example, a Google account, Apple ID, or Samsung profile. Thus, the decision on which ecosystem we operate in will have even greater consequences than before.
Communication with AI will continue to occur vocally and through chat. But increasingly, it will happen in the background and will be influenced by our context:
- what we see on the screen,
- our location,
- our current time-of-day habits,
- the meetings or anniversaries in our calendar,
- data from IoT devices, etc.
Additionally, AI will act and suggest to us unasked. Eventually, communication may come through a brain-computer interface, but for now, that is in the distant future (2035?).
Task execution will take place through a series of integrations (e.g., using various APIs), but also a special “interface” based on artificial intelligence using the Large Acton Model (LAM). Based on the analysis of a series of interfaces thanks to LAM, it will be possible to perform operations on our behalf. (An example of a device that uses LAM is, for example, Rabbit R1.) As for the actions that our digital twin could perform, just let your imagination run wild!
The Large Acton Model (LAM) is a general-purpose model using AI, understanding interfaces, and allowing operations to be performed based on them. Commands to be executed and information about their results can be processed through LLM (Large Language Model). LAM may in the future reduce the need to operate on graphical interfaces (GUI). Potentially it can operate on the real-world interfaces thanks to camera vision.
The mentioned Rabbit R1 is an example of a physical device that in subsequent versions could evolve into a digital twin. Of course, other devices and robots also can come into play. But in my opinion, the digital version has a better chance of widespread adoption – it could be an app, a layer on the phone’s operating system (a so-called launcher) or an integral part of the operating system in another form. This second approach could change the way we use phone interfaces and more.
Who knows, maybe our digital twin will also take a less serious form, such as a mascot known from old computers. Do you remember the famous paperclip, the assistant from the Microsoft Office? Such a form could serve wider adoption.
With the development of AI as a technology in itself, an e-twin could gain self-awareness and personality. It would probably be better if its character were somewhat different from ours, to avoid a strange experience. It reminds me of an episode of the series “Young Sheldon”, where the main character meets his copy from another dimension; at first, it seems like a wonderful solution (two great minds join forces), but it quickly stops working and ends in a fight…
Safety and Privacy
Since a virtual assistant, and even more so a digital twin, must know a lot about us in order to advise us and perform actions on our behalf, privacy and data protection issues become even more crucial.
To begin with, let’s be honest: it’s highly likely that corporations will know even more about us, so it’s more important to consciously make choices about whom we entrust our data. There is a good chance that the data will be anonymized, so from the perspective of a corporate employee, it will be difficult to determine exactly who they concern. However, the algorithms based on them can be terrifyingly effective. (Already at the beginning of the 21st century, personalized flyers from a certain supermarket chain in the USA “guessed” a woman’s pregnancy before she realized it herself.)
The processing of this data, and thus the direction of responses and decision-making based on data, will most likely take place on our end devices (computers, phones, etc.). This will be possible thanks to special processing units (NPUs) adapted for AI, which already exist, so over time they will have even more computing power. Powerful LLMs will be in our pocket. At the same time, more advanced operations will use computing power in the cloud, which may blur the boundaries between end devices and the internet even more.
The vision of a digital twin, which may know us better than we know ourselves (and certainly better than our closest ones) is somewhat frightening and definitely requires adjustment. In my opinion, this concept is worth treating as another argument for consciously using the internet and phone, because everything leaves some trace and can feed the algorithms.
At the same time, an e-twin can be an opportunity to unlock further layers of our potential. Perhaps the concept of a team could be redefined to encompass… one person supported by their digital twins.