THE END OF PRIVACY!? A VOTE FOR “NO.” I do not believe in catastrophic theories that assert we will have to become dependent on technology and we will not be able to control it – says Mirek Sopek from the MakoLab company in a conversation with Monika Redzisz.
Monika Redzisz: Do we really need to say goodbye to the privacy? Is there no hope?
Mirek Sopek*: Oh, there absolutely is! Technological solutions exist which will protect us from this dire vision. Profiling on the basis of what we do online does not terrify me that much. The real danger is what is happening, for example, in China today. Profiling mixed with machine learning is already leading to authentically Orwellian situations. If the parents have a poor social credit score, their children can not go to a good school. There are people who have not done anything bad, they are not criminals, but, for example, they did not take their commitments to mobile network operators all that seriously. As punishment, they cannot buy a train ticket, nor a plane ticket. There are dozens of millions of people like that! This surveillance is going too far. If I made a mistake, it is assumed I am going to make it again, my children suffer as a result. It should not be like this!
It is terrible, but it is only currently possible in a country like China.
Let us look at what is currently happening around us. Yes, we are used to democracy, GDPR protects us in Europe but some trends are alarming. Many politicians at the top of the power structure are not hiding the fact that they consider part of the achievements of liberal democracy as evil. It is beginning to be ominous. Fortunately, there are smart people who come up with technical solutions protecting our privacy. There is hope.
What solutions are they?
One of them is the Semantic Web Project, developed by Tim Berners-Lee. He created the World Wide Web in 1989; is my true guru and his new company is potentially our business partner in the near future.
Our fundamental problem today is that when we publish our data in a certain system, then that system becomes its owner.
The Web is only 31 years old, and it has already revolutionized our world. In 2004, with the creation of Facebook, we got another revolution: social media, bringing what we sometimes call Web 2.0. Then, Tim Berners-Lee lay the foundation for something entirely different: Semantic Web, a semantic network that I call a Web Full of Meanings.
How does it work?
The idea is extremely simple. It is based on swapping our existing links with “intelligent” ones. We then use what we know from logic – classes and relations – and they are defined by something we call ontology. It is not a philosophically understood ontology, of course. To put it simply, it is rather a “technical ontology.” or intelligent vocabulary. Then, we create metadata – an intelligent layer that classifies data, showing it in the context of other data, and finally, it all allows for the application of logical reasoning. Algorithms can process information better, more “intelligently,” receive information not explicitly stated but resulting e.g from the context.
It is already working in the public space. Google bought a company specializing in this field in 2013, created a gigantic data network from Wikipedia and other sources. On every well-made site there ought to be linking patterns, defining a particular person, organizing the place or a phenomenon in a contextual manner.
And how does this relate to the privacy of our data?
Our fundamental problem is that when we publish our data in a certain system, then this system becomes its owner. We cannot sue Facebook that it sells information about is to marketers because we created the account on our own, we disclosed the information about us. This is how it is organized on a technical and business level.
Tim invented a system where our data always belongs to us. He was the director of the W3C, the consortium which ensures proper standards on the Web. He is now involved in a campaign that aims at saving the web from any form of manipulation, including the political one.. He was shocked after Cambridge Analytica. He concluded it was going too far. He decided to design a platform on which our personal information will be safe (Social Linked Data).
A TOR onion-structured network already exists. Our identity is protected there.
TOR deletes all the mechanisms from network layers that make it easy to locate us. Our IP address is not recognizable but it is not something which offers data protection in itself. If someone will decide to publish some part of our data, then it will not mater if it is TOR, or not TOR, that someone will be able to steal it. Tim latest project (called Solid) works differently: here, I can have total control over what is mine. If Facebook wants to use this data, it will have to ask for permission.
It is a matter of setting up different rules?
It is something more than the rules. After all, we can never be sure the rules will be followed. It is a change on the level of technology. The entire architecture of the system changes here. Our data, pictures, movies, information, closed in special ”pods”, will be under our control. Now, when I write something on Facebook, my personal information ends up on their database servers. Facebook can disclose it to me but the data is there. Tim Berners-Lee turned this situation around. Data is saved in my pod and Facebook has to ask me to disclose it. Even when I agree, the data is still mine. The pods can freely move between various services.
Where are we going to take these pods from?
Tim’s start-up wants to create them and make them available for free.
On which level would the change happen?
Just above the level of the protocols like http – you could say that on the higher levels of network protocol. It would not be like TOR here where it happens on low levels of VPN networks. Semantic Web is strongly linked to http protocol but it also functions on higher level. Project Solid functions on the level of APIs that are also on the higher layer than http. There is an area there called Link Data Container, one can imagine it as a folder. It is our data plus a second layer – metadata which describes it; a catalogue in a form of a graph The whole data semantics (i.e. its meaning), links and also authorizations are also under our control: what do I want to give someone, what do I agree to.
Currently, we secure our data, our personal information with ciphers, encryption. When quantum computers are created, the ciphers we use today will no longer protect us.
Does that mean the entire Internet infrastructure will have to be built from scratch?
No! It can function with today’s Facebook or Amazon or other platforms. The difference will only be about who will control the data.
Is it realistic? In which technical phase is this currently in?
Still in an early phase, but intellectual luminaries stand behind it. It is a really new thing. Tim came up with the Solid idea in 2016. The company is already well founded, and he is currently looking for partners who are going to collaborate with him. Our company is also on that list, we are waiting for the information whether we have been accepted into the group. There are those who say it will never work, but I strongly believe it will.
When I share a particular piece of information about me, how will I know that he will not copy it and use it again for something only he knows? Besides, how does it pertain to clouds on which we all save our data today?
We will be able to continue to keep it there. It is simply that the data has to be properly secured. We get to cryptography here. Currently, we are securing our data, our privacy with ciphers, i.e. the encryption. When quantum computers are created, however, the ciphers we use today will no longer protect us. And here the name of an eminent Pole, comes up: Professor Artur Ekert, the director of the Centre for Quantum Technologies at the National University of Singapore. He is the co-inventor of the most important field allowing for the possibility of privacy protection – quantum cryptography. He came up with a method of data transmission of different quantum objects with the help of fiber optics which guarantees confidentiality. It is the most genius way of encryption one can imagine.
How does it work?
What is the final layer to secure the privacy? An agreement between people: when you and I agree our conversation is private. The expert invented an algorithm based on the so-called monogamous correlations. In quantum mechanics, two objects are inextricably linked with each other: when one changes state, so does the other. There is just no way to let in a third variable. Research results on the topic have been published in, among others, ‘’Nature.’’ I remember the following beautiful sentence from there: ‘’Monogamous correlations of any kind and free will are enough to protect our privacy.’’ Privacy guaranteed by quantum mechanics – isn’t it beautiful!? This securing method will be so safe that no Amazon, IBM, or Facebook will be able to decode our pods. Nobody will breach our privacy.
You deal with quantum physics?
I am a chemist-theorist and I wrote many works on quantum physics even before founding the company. I am going back to it now. For example, I take part in an interesting project where the aim is creating a blockchain capable of resisting quantum computers.
But an idea on how the quantum computers could function does not even exist yet.
We do not yet know what is the best physical carrier for the basic elements called “qubits” (quantum bits)but it is only a temporary technological problem. Importantly, this method is based on quantum physics phenomena and they work everywhere.
Let’s say that all technological problems are solved and the Solid platform is created. It is a revolutionary idea. Maybe a bit too revolutionary for those who get rich off our data today?
Obviously, this requires a change in the business model. But I am able to imagine the functioning of the model. We all need to understand the essence of this revolution. I would kindly say to pay attention to the levels of social discontent, how many there are who say: ‘’I do not want anything to do with Facebook. I quit.’’
And governments, security forces, the military? Everyone who uses surveillance of citizens for national security reasons?
There can be a problem with it, of course, but I hope there is a possibility of compromise. We need to set forth the rules that governments, companies and citizens will follow in order to ensure the Web is censorship-free while not infringing on privacy. I am a member of Data Coalition in the U.S. which is the strongest lobbying group to keep openness and maintain data privacy. The group is non-partisan.
You mentioned that you may become Tim Berners-Lee’s business partner. How did it happen? How did you end up in America?
I had loosened my ties with the university and I set up my own company in Łódź – MakoLab. We focus what is widely understood as digital services for companies. However, many years ago I removed myself from the top management and got interested in the Semantic Web. At one point, at a conference in Boston, I met a boss of the company I had worked for in Canada before. ‘’Listen, you work on Semantic Web,’’ he said to me. ‘’Let’s do something about it.’’ This is how we founded Chemical Semantics, a start-up in Gainesville, Florida, and we worked on improving data description in chemistry. We used semantic descriptions to make it easier to understand in the digital world.
So they could have their meta description?
Exactly. We became authors of an ontology which we proudly called the Gainesville Core (referring to the most serious semantic ontology – Dublin Core). At one point, my vision and that of my business partner have diverged, so I quit. I founded MakoLab in the USA some time later. In 2018, I launched another start-up which is fully owned by MakoLab. We use semantic techniques to describe important data pertaining to the companies on a global scale. Our success depends on the fate of the public system called LEI (Legal Entity Identifier), the situation of which is currently uncertain. It is about a mechanism which can be understood as a ‘’global VAT number,’’ i.e. all companies in the world are supposed to have a uniform identification code. There are more than two hundred million companies in the world, and we hoped have them all in our database in a few years. Unfortunately, the European Union partially withdrew from the idea in 2019. Indeed, the idea itself is genius, but poorly thought-out when it comes to financing, because it is financed by companies. While a hundred dollars means nothing to an American, a Polish company, it is too costly for a company in Bangladesh.
How does the Semantic Web relate to artificial intelligence?
Tim Berner-Lee’s idea does not pertain to artificial intelligence as such, but if we link machine learning methods with semantic techniques, we get something which lets us understand how ML algorithms work. Today, everyone is in awe of automatic facial recognition and voice recognition. They do not yet understand how significant semantics is for artificial intelligence. Meanwhile, the semantics lets us rebuild our faith in the fact that something which gives us a conceptual framework to interpret the world, the division of knowledge by Aristotle, something we have considered to be the canon of knowledge for two and a half thousand years – can also be used in machine learning. We can count on understanding how algorithms work thanks to semantics.
Semantics lets us rebuild our faith in that something we have considered to be the canon of knowledge for two and a half thousand years – can also be used in machine learning.
I am now driving in an autonomous vehicle and I do not know why it has stopped. What has it reacted to? Unclear. It reacted the way it had been taught to react by the algorithms. Linking them with semantics will help us understand the reason behind particular decisions. Bank algorithms reject a credit application and it is not clear why, because they do not give arguments. Or a system predicting patient health – a neural network classifies patients based on their susceptibility to schizophrenia, but on what exact basis? This we do not know.
Does that mean the unpacking of the black box?
Yes! That is why the flirt with semantics, it is the most interesting fragment of machine learning development. We will all benefit from their symbiosis. Maybe we will soon avoid a situation where algorithms indicate something and we must blindly believe it and follow it.
And I remember that in 2008 everyone in semantics disavowed AI, because it was not fashionable. It returned to good graces since then. Now. I believe everything needs time, maybe many years had to pass. Today, semantics means there is a chance for black AI boxes to gain some human intelligence.
Do you believe that the newest technologies can reverse the negative civilizational processes that we are faced with?
I do not believe in catastrophist theories that along with the development of technology we will have to become dependent on it and we will not be able to control it. I do not believe it, because there are mechanisms, belonging to the same technology which can protect us against it. But I will not try to guess if they actually will succeed. Technological solutions such as quantum cryptography or semantics have a significant potential. Scientists make many marvellous discoveries which are a blessing for us and politicians do what they have always done. Poland, for example, which has the biggest problem with environmental protection in Europe, should be the first to act. And what are we doing? Not much.
*Mirek Sopek, Ph.D.: a graduate of the Faculty of Technical Physics, Information Technology and Applied Mathematics. He earned his Ph.D. at the Faculty of Chemistry, Theoretical Chemistry .He worked as an assistant and a lecturer at Łódź University of Technology (molecular modelling), The University of Humanities and Economics in Łódź (e-commerce/cryptography) and at the Film School in Łódź (computer graphics).
In 1989, he founded MakoLab. In the years 1996-2000, he worked abroad; first, as a scientist and a programmer at HyperCube in Canada, and then as a project manager and system designer at HyperCube in the USA. He was the director of Chemical Semantics in Florida. Since 2016, he has been the director of the American division of MakoLab, and in 2018 he founded a start-up LEI.INFO operating in the field of digital company identification.
Under his supervision, the MakoLab company became a Digital Solutions Agency which specializes in digital services for the automotive sector and the financial sector. It has international branches in Paris, London and the USA.
He oversaw the creation of the Research and Development Team at MakoLab SA that deals with semantic technologies, among others. He organized international conferences three times (MakoLab Semantic Day in Paris and Warsaw) and began the MakoLab collaboration with the schema.org team (Google consortium, Bing & Yahoo) and the consortium EDMC (comprises the financial sector companies). As part of this collaboration, schema.org extensions have been created for automotive industry and the banking sector. He is the founder and the co-director of two consortium organizations W3C-GAO (automotive ontologies) and FIBO (financial ontology). Presently, he develops innovative solutions using blockchain technology (GraphChain and Quantum Blockchain).
Mirek Sopek and Michał Kosiński were guests of the last meeting: ‘’HumanTech Meetings’’ that took place on December 17, 2019 and was titled: ‘’Śladami Big Data (In the Footsteps of Big Data)”. These meetings take place periodically as part of the HumanTech Innovation Center at the SWPS University of Social Sciences and Humanities. We thank the organizers for the invitation and help in the implementation of the project.
Przeczytaj polską wersję tego tekstu TUTAJ