''Talk with the robot in a company’s call centre will be more effective than with its specialist very soon''
Big digital regulation: the legislature’s opinion about big data
Big data aren't a novelty, bodies of statistics, sociologists, companies with big clienteles use them. However, algorithms and capacity for their processing with machines have appeared not a long time ago and continue changing and developing. At a meeting organised by the Association of Russian Banks, experts discussed how AI and big data would change the banking industry and how it could be regulated by the country. Realnoe Vremya tells the details.
What is big data?
The topic for discussion was uneasy. The legislature falls behind new technologies, it doesn't have a clear definition of artificial intelligence, big data yet, there are few law cases on the topic too. At the beginning of the discussion, President of the Association of Russian Banks, corresponding member of the Russian Academy of Sciences Garegin Tosunyan expressed his opinion about the initiative to adopt a law to regulate relations between people and robots:
''We don't understand the problem completely, but we don't dare to regulate these relations. Who will be the consumer: robots or people whom we will protect from robots?''
Russian Academy of Sciences academic, Deputy Director of IT and Management federal research centre of the Russian Academy of Sciences Konstantin Rudakov agreed that the very concept ''big data'' was badly formalised. The expert offers to consider data big depending on tasks and resources for their processing. The academic believes it's obvious that summing up a million of numbers on a piece of paper is already working with big data for people, while the computer will do the task in a second.
In Rudakov's opinion, big data, as a rule, are used for very easy tasks that often come down to the question 'yes' or 'no'. The expert called data flow catastrophic and urged to use the least sufficient quantity to perform a specific task in collecting and processing. ''Data are a source that, in addition, become quickly obsolete.'' Meanwhile, very data are often in the public domain and have a low commercial value.
In Rudakov's opinion, methods of their processing and tasks they are processed for are a bigger question for potential regulation and protection. ''You can download terabytes on the Internet, but it will valuable what you can do with these data. And even the knowledge that you can do something with these data can be more valuable than the very data.''
Garegin Tosunyan: ''We don't understand the problem completely, but we don't dare to regulate these relations. Who will be the consumer: robots or people whom we will protect from robots?'' Phot: tsargrad.tv
Rise of AI is fantasists' cunning
When talking about artificial intelligence, Doctor in Physics and Mathematics, head of Information and Management federal research centre of the Russian Academy of Sciences Mikhail Zabezhailo offered to consider it as an assistant, not a system for making decisions. ''A person will be responsible for these decisions anyway,'' the expert noted.
Zabezhailo also mentioned the vagueness of formulations of the question. What people usually call artificial intelligence now is machine learning systems, in fact. They don't comply with definitions of AI that was formulated in the middle of the last century.
''What will we regulate? We constantly hear an incantation on TV that a rise of the robots is about to come and artificial intelligence will kill everyone. From a perspective of a professional, there is some cunning here. AI is a kind of device that a person programmes and these programmes are fully understandable and clear for a user. ''The black box isn't allowed,'' Mikhail Zabezhailo explained.
Big personal data
The Internet has a big number of personal data, the access to them is not limited – there are social networks. Banks have already learnt to use them by processing a mass of information. One can make such a detailed portrait of a client by likes and reposts, so the bank can decide whether to grant you a loan or deny. For instance, Sberbank analyses borrowers this way together with a classic evaluation of creditworthiness.
Doctor of Juridical Sciences, President of IP CLUB, Professor of the Kutafin Moscow State Law University Marina Rozhkova offered to distinguish very data and their use: ''When we talk about big data, we talk about not the source but how they will be used. From a perspective of law, it is a value.''
The spheres where data are used are numerous: insurance, commerce, medicine. In addition, the jurist noted that one and the same data could be used in completely different ways. ''If some information appeared 2000 years ago, it was used for its direct purpose. Today we can get big data, change them with the help of different algorithms, only just for your personal goals.''
Rozhkova told about a law case when data of Vk.com social network were used in the National Credit Reporting Agency to create a portrait of their potential borrower. The social network thought it had the exclusive right to possess the users' data and filed a lawsuit against the agency. The case ended with a peace agreement. Not recognising the complaint, the agency stated it wouldn't use the data by violating the legislature. Meanwhile, the court noted that all companies could use data that are in the public domain. During the discussion, Mikhail Zabezhailo paid attention that even if somebody's right of possession to big data is recognised, it will be very difficult and expensive to protect them. Unlike their analysing methods and algorithm, this issue still needs to be studied.
Deputy Director General of SAP in CIS countries Dmitry Krasyukov noted that now so-called 'taps' are used in data centres because few people understand what to do with big data. And the volume of information grows and requires development of AI and machine learning systems. Krasyukov is sure that a talk with a robot in a company's call centre will be more effective than a talk with its specialist very soon.
What needs to be regulated?
Discussions about the work with information, protection of personal data and problems of big data are only arising. Developed machine learning systems already can make unusual decisions from a mass. For instance, loan denial because the robot did not like reposts. However, participants of the discussion agree, in general, that there is no sense in trying to protect very big data from theft or use of by third parties. Probably processing methods and analysis tasks of very data will be regulated when inert legal systems of countries reach it.
The speakers could not make up how relations between people and robots in the legal framework would be regulated. However, in Mikhail Zbezhailo's opinion, there won't be problems with technically correct artificial intelligence, at least, until the system doesn't make decisions itself but just complements a specialist's intelligence.