The next decade will be characterized by a total paradigm shift in the interaction between man and machine and this will take place along two parallel and complementary directions: the use of the voice and the massive adoption of Extended Reality.
What we are doing today, a little shyly and sometimes feeling a little stupid, with voice assistants like Siri and Alexa, will become the standard, and sometimes the only way, we will relate to the service providers we need; not therefore only basic home automation, but a true and complete interaction with heterogeneous external systems and service providers of all kinds. We will therefore shift from the current “Alexa, turn on the kitchen light” to a much more interesting “Hey Siri, call the plumber and arrange an appointment for one morning next week, but only on a day when it rains because if it is nice I want to go for a run in the park”, or from an ordinary “Ok Google, what’s the weather forecast for tomorrow?” to a more useful “Alexa, book the conference room and organize the videoconference for the Monday morning meeting, send an invitation and agenda to everyone and set the coffee break for 10.30. Don’t forget the gluten-free croissants and the decaf. The cost of the conference room goes on Concetta’s cost center”.
The level of interaction will quickly become comparable to what we have between human beings, as an improved understanding of the context will be added to the improvement of the natural language recognition component, and above all to the need to transform complex sentences into operational sequences of commands to be put in the right order to be transmitted to external service providers which will be part of the ecosystem.
Those who provide services will therefore also have to ensure they are interoperable and integrated within that ecosystem, so that they may no longer be used directly by human action alone, but also and mainly in “machine to machine” mode, i.e. directly by other machines which will relate with them fully automatically.
From Reality to Extended Reality
The second item will instead be enabled by a new generation of devices of which we have already seen a few prototypes in the past and of which we are now starting to see the first examples, slightly limited for technological reasons and because they lack an actual ecosystem of content and services: these are the smart glasses, glasses able to improve our interaction with reality. Actually, they are not new devices, we have already seen Google Glasses in the past and today there are also more recent versions on the market, such as the Bosh and Epson ones. However, the devices that we find on the market today will not be the real factors of change, there is a need for a significant qualitative leap in terms of technology, infrastructure, ecosystem, business model and availability of applications and services.
These will be devices able to make us fully experience what is called Extended Reality, the union of different paradigms of interaction with reality:
- Virtual and partially immersive Reality: the lenses will in all respects become non-transparent screens on which to view video content, typical virtual reality experiences, videogames, but also new ways of training, virtual tourism and new experiences which are still to be identified.
- Augmented Reality: the lenses will become transparent and will add new information and content layers directly superimposed on the traditional vision, this will allow you to drive with the navigator’s instructions directly in the field of vision, to automatically access information about a place you are visiting, a work of art that you are looking at in a museum or a person you are talking to. When an apparently unknown person stops us on the street and greets us, we will be able to know who he/she is, if we know him or her, what kind of relationship we have and on what occasion we met him or her. Looking at a monument in a city we will obtain historical, tourist or other information, whether free of charge or by paying, thanks to the simple fact of having a screen in front of our eyes which is capable of superimposing a layer of digital data on reality.
- Mixed Reality: an evolution of augmented reality, in which we are able not only to access information layers superimposed on the traditional vision of physical reality, but also to interact directly with these by using our voice or by specific gestures. When visiting a museum, for example, we will be able to buy the entrance ticket, make the payment, access the map of the museum site, select the path we wish to follow and use a complete smart guide to the pieces. While we shop, we will be able to have the shopping list available, to automatically cross out products by simply having scanned their bar code with the glasses, and to pay directly from the glasses, without even stopping at the checkout.
In this case too, there will be a lot of work needed by the application providers who will find a completely new channel through which to deliver their content and services and, given the immediacy of interaction that these devices will provide, it is conceivable that this channel will, in the nearer future, become a priority one. We could therefore move from “mobile first” to “glasses first”.
The availability of these new technologies will naturally introduce new levels of complexity in privacy management and in the necessary awareness which we will need to have when managing our data and when transferring these to third parties. The fact of having a camera and a microphone always switched on could lead to some initial difficulties in relationships with our interlocutors, but it will only be a matter of time and everyone will get used to it, exactly as we have become used to seeing people on the street who apparently speak to themselves, but in reality are using earphones and headphones to communicate with others; once this behavior was somewhat strange, now we no longer notice.
A new role for Artificial Intelligence
These interaction technologies will need to be supported by robust, modern and efficient back-end infrastructures and in this context a great role will be occupied by Artificial Intelligence algorithms which will feed on the data that we will constantly produce.
Our glasses and our virtual assistants, together with all the other devices which we own and that will be part of a single large ecosystem, will learn from our behaviors, will be able to offer us precisely what we need at that moment increasingly more often and better, from the bus ticket when we are at the bus stop, to reviews of a product we are looking at, to the best price found on the Internet for that product.
Where data will increasingly be the fuel of our existence, algorithms will represent the engine capable of using them and of transforming them into value and this value will be made available to us through our smart glasses and virtual assistants.
One of the most interesting aspects of the Artificial Intelligence algorithms of the future will be the concrete possibility of avoiding BIASES, that is, the algorithmic biases which inevitably pollute the machine’s experience during the learning phases and which therefore jeopardize its proper operation. As we ourselves are conditioned by experience, the same happens for machines which learn from us, with the result that our prejudices, whether conscious or unconscious, will inevitably be introduced in the knowledge base used by the algorithms and will consequently reflect on their behaviors. If when we go out for dinner we usually go to Japanese restaurants because we like sushi, the algorithm will learn that our habit is that one, that we prefer that type of cuisine and therefore will almost exclusively suggest Japanese restaurants, this is a good thing as the algorithm can serve us better if it best meets our desires and needs. However, if we work in an HR department and the algorithm realizes that candidates of a certain ethnic group are inevitably rejected regardless of the other characteristics of their professional and personal profile, they will learn that that ethnicity is a negative element and that those workers must be penalized during the selection phase. It will therefore behave in a racist way because it will adapt its behavior to what it has been taught, incorporating all its users’ BIASES into its knowledge base, both preferences (which is what we want) and racist prejudices (which is not what we want).
Numerous strategies are being studied to overcome this problem and within a few years we will have available algorithms capable of absorbing BIASES, considering them a phenomenon to be controlled and not be subjected to; they will therefore be able to discriminate between the various pieces of information and understand which are to be considered part of the knowledge base and which ones can be discarded. We will thus be dealing with algorithms capable of acting fairly, justly and respectfully towards racial and gender differences, therefore, in a certain sense, and from this point of view, algorithms which are better than many human beings.
The relationship between artificial intelligence and blockchain
Another interesting aspect of future Artificial Intelligence algorithms will be represented by the relationships between these and the various incarnations of the Blockchain technologies. Today these technologies have, for the most part, the characteristic of being decentralized systems (and not just distributed), safe, transparent and essentially unchangeable.
In the future, some of these technologies could represent an extremely safe way to store information, ensuring its immutability and security over time; we are not of course talking about the current versions of the Blockchains which host cryptocurrencies, but about their potential evolutions which will allow to store some information directly on-chain without the limitations of the current platforms.
With these characteristics, they could become privileged sources of data to be given as input to Artificial Intelligence algorithms and therefore could be part of the knowledge base which will determine the behavior of the algorithms themselves.
In the same way, Artificial Intelligence algorithms will most likely flank smart contracts, those portions of traditional code which work, under certain conditions, directly on the data present on the Blockchain. These new algorithms may be applied to the data on the Blockchain and they may perform complex operations on these which go far beyond what is now allowed by traditional smart contracts.
In parallel, the distributed and decentralized computational power offered by Blockchain nodes can be made available for executing on-chain Artificial Intelligence algorithms. These are computational resources usually dedicated to cryptocurrency mining, but which, if necessary and on request, can be diverted at least partially to this type of processing.
The presence of these algorithms on the Blockchain will make them immediately usable by anyone, therefore a new market will open up, that of Open Algorithms which can also be delivered in Algorithms As A Service mode.
What companies need to do over the next ten years
Companies that want to seize the new opportunities enabled by these new paradigms will have to start thinking in terms of ecosystem and real value which can be brought to the consumer and the citizen thanks also to the enabling factor which technologies as a whole can provide. However, it is important not to focus on the single technology as the real added value will be provided mainly by the ability to integrate existing technologies, the processes which need to be rethought and the innovative drive of emerging technologies.
The fruition channels will change radically, we will habitually converse with voice assistants and we will see the world through the intermediation of glasses capable of providing us with content and services in new ways. Companies and the Public Administration will therefore have to make their services part of a single integrated ecosystem, so that these services can be used in completely new ways, which we are not yet used to today. The game will be played in the field of integration, interoperability and of new models of human-machine interaction.
These new AI algorithms will think of ways to further simplify our relationship with these contents and services, as they are also part of the ecosystem and finally cleansed of some of the great technological and behavioral limits which today still limit their diffusion and use. In this case, the key to everything will be the quantity and quality of the data available and, above all, the ability that companies will have in identifying new correlations between data apparently without any connection. Where the datum in itself is a value, the correlation between data represents a superior value.
Therefore, a new way of interacting with reality awaits us, enabled by new technologies and by important evolutions of the existing technologies, but with great attention, at all levels, because of the issue of privacy and of the management of personal data.
Massimo Canducci
