The infusion of digital technologies into the basic ground of our society is the biggest transformation phenomena characterizing our post-industrial society during last 20 years or so. There is a clear tendency that is asking us to put machines, and the overall mechanization, at the heart of every process transformation in every industry sector.
The ingredients that allow all of that are well known and include the pervasive capacity to interconnect things in networks and collect data from the physical world (Internet of Things), the wide availability and the ability to process huge quantity of digital data produced by humans and by machines (Big Data, Social) and the availability of more effective tools for data analysis to learn interesting phenomena from data collected (Analytics, Artificial Intelligence).
The human-machine relation is profoundly changing aiming to be as natural as possible.
All those technologies are following a “natural evolution path”, we could say with one sentence, that is aiming to add continuous values and improve each of them along various dimensions such as better performance, new efficiency, completeness of functions, new functions and so on.
In addition to this “natural evolution path” more recently, a new emerging trajectory can be equally observed that considers all those technologies and perhaps others as a sort of elements of a larger equation. In other words they can be merged, combined to create not only new technologies but completely new things (and new needs someone could argue).
I use to refer to this area Active Intelligence. What an Active Intelligent System?
Well, If you consider just last 20 years or so, we learned to see computers as valid tools able to help us during our business as well as during our personal lives. Obviously one of the major contribute was coming from the introduction of Internet itself in our life and by the possibility to accumulate and search in an a huge quantity of data. We started, more recently, to recognize not only this foundation role and value of the so call digital transformation but we understood that what we are accumulating by each of us, as a group of ants, can provide a starting point to generate even an additional value.
More or less we started to label this phenomena in overall as a data-economy and data-driven society saying that the technology drivers of this transformation are Big Data, Internet of Things, Cloud, Mobile, Analytics and others let us consider digital data as the “new oil”.
This was/is/will be the first step. Let’s go ahead now. 🙂
We learned that once understood the digital data can provide even an additional value with respect the original content, so here came out a large part of the new spring of Artificial Intelligence & co more recently. This last is the key ingredient to transform “data points” on only into other data but into “decision points” targeting all interested parties.
So we started to think how we can transform full businesses in new terms: “Powered by AI” is actually and right now the new slogan and mantra for many.
Artificial Intelligence, as a set of computational methods, is materialized in business terms with a number of names such as Cognitive Computing, Augmented Intelligence, Smart Machines or simply Machine Learning (for Business) and perhaps others.
The baseline is inherently the transformation of the computers from simple tools that have a complete “passive role”, more or less corresponding to the ability to manage a huge archive of data, into tools that could support someone, for professional reasons or not, to understand the reason for which he/she is searching into that archive of data.
We have now plenty of examples of systems that provide such a kind of abilities. Also their ability to communicate with humans improved a lot and we are starting to think to a computer not only in terms of keyboards, monitors or mouses rather into completely new kind of tools and devices.
Now we are entering into a new critical evolution point that I personally call “decision lakes“.
What are decision lakes?
Well they are a similar phenomena to what Big Data did with Data.
In other terms the AI & co is helping us to make a decision in a number of contexts and applications but at the same time it is exposing us to a wider solution searching space that perhaps includes decisions that we were even not able to figure out, easily or quickly at least, without an artificial help or artificial prosthesis.
A couple of examples to explain what I mean for Decisions Lakes are here.
Example 1: Consider the cooking context. There are funny AI systems that have been proposed to help passionate and professionals to move into the ingredient combinatorial space to invent and propose, for instance, novel recipes that combine and mix a certain number of ingredients as well as provide a reasonable novel cooking procedure. An example of such systems is the well known Watson Chef (https://www.ibmchefwatson.com/). Well, the largest cooking internet site stores a recipe collection whose size is about 10 power 6, on the other side, some researchers estimate the number of possible combination of ingredients to look for novel recipes in something equivalent to 10 power 23 and even more. A decision lake is exactly here in the difference between 10 power 6 to 10 power 23 where the AI system explores and provides solution hypothesis to consider and on the other side, the human, is in a paradoxically state to have more options to digest (it is the case to say this for this example!).
Example 2: Another example could come from the Affective Computing or as someone use to define Emotional AI area.
Here we can recognize the presence of a similar “decisions lake” situation. In fact, an Affective Computing AI algorithm could interpret signals about emotional aspects of an individual in a much more wider ways than an human use to do, at least in an explicit way, or at least this is the claim 🙂
Let me provide a simple example also for this case considering a tool from the IBM AI tool box. If you consider Watson Personality Insight (https://www.ibm.com/watson/services/personality-insights/) a system that is able by applying psycho-linguistic techniques and models to derive a reasonable profile about who wrote the text. This system generates a profile that includes more than 50 characteristics referred to the personality, needs and values. It leverages personality models, human behavior models that individuate which aspects of a product are likely to resonate with the author of the text and possible human values that are the motivating factors that influence the author’s decision-making, Here we have another potential decisions lake that is in difference between the quantity of traits that reasonably we build naturally in our minds, in our case could be as simple as “I like this guy” or “I do not like him” that
are the tip of an iceberg we have implicitly in our minds, and an overloading quantity of details about the same topic that this tool could generates
and revers under your eyes.
I selected these two extreme examples to explain, from my point of view, the concept of the Decision Lake but now if you think about it. I am quite sure, you
will find plenty of Decisions Lakes that are forming as a result of whatever kind of AI technology and its applications you will consider.
What is the practical result of this trajectory?
We are entering into a new step of the computing evolution, inexorably, that is driven by the combination of technologies to which I was referring at the beginning of this article and the creation of what I am referring here as Active Intelligent Systems.
As a consequence the tendency in IT to design things will be opposite of what we did so far: instead to think to a computer as a tool that helps to solve a problem we will think more and more in terms of completely new products and systems to invent.
They will be not simply Smart Products or Smart Objects that we obtain by adding a digital ingredient to transitional objects, rather they will be new products that aim and create new needs.
This figure represents this trajectory I am trying to describe that has a natural end point: thanks to the combination of the IT technologies I mentioned before we are creating de facto New Products and New Markets.
From a technical point of view, this combination of technology forces is enabling the creation of a new generation of systems that not only are able to learn new tasks but also have autonomous capabilities, presenting proactive behaviors and stimulation with respect to the environment or the process in which they are placed.
An Active Intelligent System can help us to cope with the problem we have to deal with Decision Lakes and the decision overloading.
Systems of this class are no more “passive” computational entities, rather they actively and pro-actively solve problems and tasks for us as well as, most importantly, potentially mediate us.
An Active Intelligent System aims to have a deep reconstruction of the environment or the process in which it is placed, as well as a digital reconstruction of the actors that work and operate in that environment. The distinguishing feature is the ability of the system to observe and learn complex behaviors from actors operating in the environment and, at the same time, autonomously activate actions and provide knowledge in a proactive and selective way to improve the overall efficiency of the environment itself, considering specific target variables.
We have already early examples of Active Intelligence Systems.
A nice and funny example of such a kind of systems is the Cognitive Dress designed by Marchesa and IBM a couple of years ago and now exposed into the Henry Ford museum (https://www.thehenryford.org/collections-and-research/digital-collections/artifact/438941).
The Cognitive Dress, in the essence, is a system-dress and it is quite simple: it is a dress covered by a number of LEDs that can change colors. The system-dress analyses the emotional mood expressed by people that are looking at the dress, for example reading real time comments, and react to the comments changing colors of the LEDs according to the mood it interprets. The objective is to satisfy or increase the emotional impact of that dress with respect who is watching it.
The model or who is wearing the system-dress is mediated, in such as sense, by this Active Intelligent System. She cannot change the colors of the dress, the dress autonomously is reacting to the environment trying to improve the look of the dress owner, following its original directives.
Of course this is a funny exercise and an extreme example but I think it is able to explain what I mean, to be, for some extend “mediated” by an Active Intelligent System.
So, in summary, Active Intelligent System could represent last in order of time of a possible evolution chain of computing systems: from tool, to butler, to concierge, to assistant, to trainer, to coach and finally to a mediator.
So, at the end of the day, you can intersect now the Active Intelligence concept I described above with all other key entities and decision lakes that are forming thanks to AI & co around them, in all kind of business and in our lives and I am sure new products and systems will be invented.
I suspect we are going to see at the horizon new Active Intelligent factories, new Active Intelligent assembly/production lines, new Active Intelligent products, new Active Intelligent materials, new Active Intelligent processes and so on.