October 26, 2020 § Leave a comment
The idea of a search engine devoted to finding ongoing behaviors may sound speculative. However, some recent developments suggest that it could become a reality within a few years. To start with, Gartner has included the Internet of Behaviors (IoB) as one of the Top Strategic Technology Trends for 2021and suggests it will touch almost half of the population. This is not surprising since IoB – analogous to IoT and even IoE (Internet of Everyhting) – has been one of the strongest trends in internet services for a decade, although the concept IoB has not been used so far. When I introduced it nine years ago in the gotepoem blog and after that in several writings, including my book On The Edge of Human technology – An Essay, it was received with puzzled looks. Nobody among the operators and technology investors I met knew why invest in it.
Gartner vs the human form of IoB
I want to make clear the difference between ‘my’ human IoB and what Gartner now suggests. The human IoB is aimed at protecting our identities by separating behavior and identity data whereas the technology orientated IoB described by Gartner, is meant to link a person digitally to their actions. Gartner’s definition keeps the person and her behavior as a useful data entity, and it raises questions. Of course, both human and technological approaches will live in parallel but the privacy and security requirements remain.
EU Parliament has recently taken actions to introduce serious limitations on the use of human and behavior data, like in behavioral ads and in microtargeting and on the related algorithms by the famous data giants. The human version of IoB can offer a Plan B in case hard decisions by EU make the life of the familiar giants, service providers and campaigners difficult. This is where the new search engine concept – I like to call it Birch here – could find a place. Birch is a symbolic and practical tree with many cultural meanings for us Finns, from saunas to traditional landscapes and it is beautiful material for Finnish design and architecture. For a Finn, the birch is a resilient tree and its leaves have healthy properties. Hence the symbolism in this so far speculative idea.
The secret of the human IoB is to find and share accurate and timely knowledge of ongoing behaviors, and to serve the behaving people (or organizations and why not machines) without identity knowledge. It is of huge value – when used wisely and with the support of a suitable data models and communication architecture. I will not go into technical details here but will only introduce the conceptual idea of the IoB search.
The nature of behavior data must be reconsidered
IoB, in its present forms, is a major business force in the Internet. My educated guess is that the amount of ‘human data’ in the net increases exponentially and will continue to do so in the foreseeable future. IoB features occur in various apps, trackers and services using human behavior data, either physiological or motor, or even symbolic like in Spotify where the user expresses her preferences to choose a suitable music style or genre. The data giants do their best to share and multiply behavior data. IoB related apps do not typically have a common data architecture and their technological solutions vary and a mature human IoB architecture could change this and allow totally new services based on the knowledge on ongoing mental and physical behaviors.
Human IoB utilizes an architecture and behavior addressing models where ‘targeting’ of ongoing behaviors is possible in any human – mental and physical activities like in sports, entertainment, health care, education, and science work, you name it.
What could Birch be?
In my forthcoming novel Perceptions of the Les Demoiselles d’Avignon the protagonist is a young theoretical physicist, Johan Ek, curious about a mathematician, Maurice Princet who in the early 1900’s introduced 4D geometry to Picasso, who was preparing his masterwork of cubism. Johan wanted to know if anyone on earth, at that moment, was working on this very specific topic, the history of Princet of whom there was only scarce information available and photographs of him were difficult to find. Johan used the Internet of Science Behavior service:
“Marianne nodded at Johan, pointing at the screen showing a mosaic of icons again. ’There you have the people in our network who, right now, at this very moment, have something going on related to Princet. I could show you how these people are related—if they are—but this is not the point here now. On the right you can see the list of IoBs codes or addresses they have decided to use and under which they work. The first code is Maurice Princet’s divorce, meaning someone is researching that—it is not gossip journalism, these are serious people—then there is Maurice Princet and the fourth dimension in art, then Maurice Princet and the painting Les Demoiselles, and Maurice Princet’s life after the Bateau.’ ”
Birch needs an architecture and a communication model of its own. The basic net architecture is already there, but what is missing, is a standard model for representing behavior data within this architecture and a communication model which makes it possible for IoB users to remain unidentified while having access to whatever the service providers and apps offer for them to choose.
An evolutionary development of a Birch-like search paradigm is possible by starting from any application sector, where massive behavior data already exists and is collected but where the separation between identity and behavior data is not inbuilt. We did a proof-of-concept app with the sw designed by Muzaffer Topdagi, a colleague and a friend, and built a situationally intelligent IoB-kind of a situational intelligence model to be used by an unlimited number of unknown own users.
The overall platform of IoB where Birch could work
Trying to introduce the business and service value of IoB (emphasizing the separation between behavior and identity data) I noticed that it was not self-evident at all to the mindset of the people I approached, what the business models and the technology for IoB could be. Birch is no exception to this, I assume. Current practices and data paradigms have adapted us to the tight bond between the behavior and identity of a ‘target person’. In my book I have described a few simple examples, but the basic ‘IoB design principles’ are the following:
- IoB is a means for people to indicate any of their ongoing, mental or physical behaviors. It can be of any form of human act or thought, from practical to abstract. The scope of behaviors is unlimited. These ongoing behaviors can then be searched.
- Each behavior is linked with a code or an address that allows ‘digital or other (!) targeting’ of the occurring behavior. The knowledge of these data is shared between people and organizations using the IoB and the app/service providers. The codes can be public, private and even context specific. This provides the possibility for both general and extreme4ly specific searches.
- A communication architecture is used by the apps and service providers so that they can detect the occurrence of a behavior or several instances (e.g. globally) of that specific behavior and to use selected communication channels for presenting their offerings. The architecture serves both global and local behaviors.
- A communication platform allows the users of IoB, wherever they are, to receive offers and services without revealing their identities or locations: they can react to or choose the ones relevant and interesting for them. (Note: because the offerings are provided as a reaction to the detection of the ongoing behaviors, there can be a very high degree of relevance in them, often 100% accurate. Service providers that don’t perform well can be switched off or neglected by the users.)
- Hubs can be formed within IoB, making it possible to integrate several service providers, each with their own commercial or public profile and offerings. This is important since often people want services or offers that have many relevant contributors at the same time but which usually do not operate together or even in synchrony.
- … and some more.
Finally, what could the present search engine providers do? They could take a step close to what Facebook, for example, is already doing, but could significantly boost it with the IoB. The fastest way ahead is a pilot, where the standard search engine is supplemented with a specific, added human IoB feature. I have described something like that (a slightly different perspective though, to improve search relevance) in my book. A simple pilot would not immediately solve the problem of separating identity data and behavior data, and some additional and tailored work is needed for that. Further, specific tools are required for recognizing what in the present searched data is about ongoing behavior and what is not. In the human IoB this is always known.
AI and IoB
At present, an AI based solution can look at the objective and subjective timing (minutes, hours, days, weeks, future intentions, plans) and other informative aspects of the behavior data sources and analyze the text and other content to reveal whether the data concerns an ongoing, intended, or planned behavior. While this is not a way to build the human IoB it would teach a lot and make possible to test various IoB features and benefits. It would teach people to look for and express ongoing behaviors, which they already now do in Facebook and elsewhere in a crude form, but they could do it in a purposeful and secure way. Service providers, global players especially, could use these experiences to avoid the apparent damage caused by the new Internet restrictions on behavior ads and microtargeting.
The ethical aim of human IoB is to keep the behavior and identity data separate. This is crucial. When AI is combined with human IoB a ‘sandwich model’ of IoB services could emerge. Its main function would be to let users express their behaviors with as simple tools as possible, sometimes supported by a learning AI. This would increase the situational quality of the behavior data that comes from honest subjective expressions. On the other hand, AI can learn to find relevant services and service providers, which best fit with the needs related to the expressed behaviors.
The sandwich model leads to an improved situational intelligence of services. It is much better than is typical today when machine learning based, statistical guessing dominates and causes continuous, daily and hourly disturbances to people.